This technology models the strain-independent hysteresis of piezo actuators by dividing it into monotonically increasing and decreasing input voltage branches. It constructs the necessary first-order minor loop datasets for each branch using an Adaptive Neuro-Fuzzy Inference System (ANFIS) for Congruence-Based Hysteresis (CBH) modeling and compensation.
While piezo stack actuators are used in vibration control, nanopositioning, and robotics due to their fast response and high energy density, their industrial application is limited by hysteresis nonlinearity, where output depends on both current input and past history, leading to positioning errors. Existing CBH models approximate minor loop datasets with polynomials, limiting accuracy for irregular and complex hysteresis.
This technology uses only two variables—the instantaneous input value and the starting voltage of the first-order curve—as ANFIS inputs to estimate minor loop values via a Sugeno-type fuzzy evaluator. It constructs an inverse model dataset in the same way to create a compensator that cancels out hysteresis effects. Applicable to semiconductor lithography stages, nanopositioners, and piezo valve drivers, it achieves higher precision in open-loop position control than polynomial models.
N/A
