Nonlinear Dynamics Modeling of the Vibrating Single-element Transducer for 3D Ultrasound Image Reconstruction using Recurrent Neural Network

In our previous work, 3D high-frequency ultrasound imaging using a vibrating single-element transducer driven by a tubular piezoelectric actuator was proved to be feasible. Considering the inoperability of imaging opaque phantom using imaging system containing optical positioning module (PSD), the acquisition of the transducer trajectory is a crucial problem for image reconstruction. The vibrating transducer is regarded as a base-excited cantilever with nonlinear effects. However, the accuracy of the established nonlinear system model declines due to the change of the working environment. In this work, recurrent neural network (RNN) is adopted to identify the nonlinear dynamics model of the cantilevered transducer to predict the accurate trajectory of the transducer for volume reconstruction.