Predictor-based Equalization of Magnetoelectric Sensor with Integrated Coil

Abstract

Magnetoelectric (ME) sensors provide high sensitivity at resonance, but exhibit limited usable bandwidth. This work proposes a prediction-based equalization approach to extend the usable bandwidth using broadband noise excitation and linear prediction. A preprocessing pipeline and spectral regularization ensure stable filter design. The resulting equalizer flattens the sensor response around resonance and significantly increases the effective bandwidth, at the expense of reduced peak sensitivity and signal-to-noise ratio (SNR) near resonance.