Intelligent ultrasonic signal processing for solving inverse scattering problems
Wolfgang H. Sachse, Igor Grabec · The Journal of the Acoustical Society of America · 1991
This paper describes new intelligent or neural-like processing procedures by which the ultrasonic signals scattered by an obstacle can be processed to recover the characteristics of the unknown scatterer. In particular, it focuses on the recently developed automatic modeler by which the characteristics of the scattering phenomenon from a number of scatterers is learned using a set of systematic learning data. It is demonstrated that ultrasonic signals measured in subsequent tests from an unknown scatterer can be processed to optimally recover the characteristics of the unknown obstacle. The solution of the inverse scattering problem that is obtained is thus completely empirical and not based on any elastodynamic theory. [Work supported by the National Science Foundation under Grant MSM-8904384.]