PARAMETRIC MODELING AND ESTIMATION OF ULTRASONIC FIELDS USING A SYSTEM-IDENTIFICATION TECHNIQUE

A system identification technique is presented for the parametric mode ling and estimation of acoustic fields generated by an ultrasonic tran sducer. In this technique, the ultrasonic fields are modeled by the su perposition of Gaussian beams. The model order, i.e., the number of th e Gaussian beams used, and the characteristic coefficients of these Ga ussian beams for the superposition depend on the source distribution. An experimental system is built to measure the acoustic fields. From t he measured amplitude and phase data, a maximum likelihood estimator i s utilized to determine the order and characteristic coefficients of t he field model. Once the order and coefficients of the field model are estimated, a complete parametric description of the acoustic fields i s then obtained. Using the estimated model, the prediction of the acou stic fields can be made in both forward and backward direction. The pr edicted and measured fields are then compared to validate the model, T he experimental results demonstrate the capability of the technique fo r the parametric description of ultrasonic fields. (C) 1996 Acoustical Society of America.