Authors: R. Khalilyulin, G. Schrag, G. Wachutka
Affilation: Munich University of Technology, Germany
Pages: 581 - 584
Keywords: system-level modeling, paramter extraction, calibration, accelerometer
We present an elaborated parameter extraction and model validation procedure for high-level transducer models that are aimed for system-level design as well as for hardware-software co-simulation (Matlab/Simulink transducer model embedded in control circuitry) in order to optimize the operation of the entire sensor systems with respect to the specific application area. Nowadays, such “smart” sensors systems comprise functionalities for the compensation of environmental impacts like control concepts, signal conditioning, and self test, e.g., and, hence, cannot be designed adequately without the help of accurate high-level transducer models, which describe correctly all imperfections originating from the manufacturing process and all impacts caused by variations of the environmental conditions. We derive such predictive simulation models by following a systematic calibration and validation procedure, starting with high-fidelity continuous-field simulations to determine all relevant model parameters and their dependencies on external impacts, followed by a proper experimental calibration procedure in order to adjust the deviations caused by manufacturing tolerances and fluctuations. Static and dynamic characterization has been carried out using a white-light interferometer in combination with a vacuum chamber and an HP impedance analyzer. System-level simulations demonstrate the necessity of predictive and carefully calibrated reliable transducer models.