Skip to Main Content
In this paper, a systematic design based on the robust control theory is developed for a microelectromechanical systems nanopositioning/probing device. The device is fabricated on a silicon-on-insulator substrate, and provides decoupled XY motion by using a parallel kinematics mechanism design. Each axis of the device is actuated by linear comb-drives and the corresponding displacements are sensed by separate comb structures. To improve the sensing resolution, the sensing and driving combs are electrically isolated. The nonlinear dynamic model between the actuation voltage and the sensed displacement will increase the complexity of model identification and control design. We circumvent the nonlinear model by redefining the input and output (I/O) signals during the model definition and identification, which results in linear and time-invariant models. A dynamical model of the system is identified through experimental input-output frequency-domain identification. The implemented H∞ control design achieves a significant improvement over the response speed, where the bandwidths from the closed-loop sensitivity and complementary sensitivity functions, respectively, are 68 and 74 Hz. When compared to open-loop characteristics, enhancement in reliability and repeatability (robustness to uncertainties) as well as noise attenuation (by over 12%) is demonstrated through this design.