Nanopositioning With Multiple Sensors: A Case Study in Data Storage

The use of multiple sensors for nanopositioning is well motivated in high-density data-storage devices, such as probe-based data storage devices. In these devices, ultra-high positioning accuracies are desired over their lifetime in the presence of external disturbances. In addition to meeting the stringent requirements for nanopositioning, the use of multiple sensors lends itself for performing sensor fusion and obtaining reliable estimates of position and velocity. These estimates can be used for the detection of shocks and vibrations and the calibration of sensors. The objective of this work is to provide a comprehensive overview of the problem of nanopositioning using multiple sensors in the context of probe-based data storage. We present the control of a micro-scanner used in a probe-based data-storage device with two sensors, a global thermal position sensor and medium-derived positional information. In one approach, Kalman and
H
_∞
filtering are employed to perform sensor fusion, and the resulting state estimates in turn serve to design feedback controllers. An alternate approach is to design a multiple-input single-output controller using either the
H
₂
or
H
_∞
control paradigm. These controllers are investigated in detail to reveal the underlying structure and the inherent sensor fusion that takes place. Comparisons are made between the two approaches.