System identification and control of a miniature external mechanical vibration device towards clinical ultrasound shear wave elastography

Abstract

An acoustic radiation force (ARF) is commonly used to generate shear waves in ultrasound systems for shear wave elastography (SWE). However, ARF requires complex hardware that produces thermal stresses in tissues and electronic components. External mechanical vibration (EMV) is less power-exhaustive compared to ARF. Additionally, EMV-based shear waves have been shown to produce larger, local displacements in comparison to ARF-based shear waves which is beneficial for imaging deeper tissues. EMV provides the opportunity for the development of a low-cost, compact, and efficient alternative to ARF-based SWE systems. A miniature EMV SWE system consisting of up to two vibrating voice coil actuators (VCAs) attached to a commercial ultrasound probe was previously designed and developed. The vibration of the VCAs was synchronized with the commercial ultrasound system to induce shear waves, replacing the ARF.

Preliminary testing on this system has shown the system's ability to produce results comparable to ARF-based SWE. In this work, the miniature EMV SWE system is further developed. The VCA is replaced with another that has an integrated position sensor. This eliminates complex external sensing circuitry. The circuitry is simplified and placed on a PCB, further decreasing the footprint of the system while increasing the robustness of the electronic connections. The system is modeled and experimental data is collected to validate the model. Data is collected using a frequency sweep to obtain the magnitude and phase of the signals of interest. A lead controller is designed to perform position control on the VCA, and its performance is evaluated. The bode plot of the system model is found to have no phase margin at the crossover frequency, therefore a lead controller is designed to provide a phase boost. The controller successfully provides a phase boost.

However, the experiments reveal the difference between theoretical models and physically feasible results. The VCA exhibits non-linear behavior at low frequencies possibly due to friction. The VCA also has a maximum vibration frequency between 80-90 Hz. The phase margin increases as intended before this limit but rapidly declines as the system approaches this limit. Further exploration into the VCA is necessary, and additional studies are needed to validate the system's efficacy on phantoms, ex-vivo tissues, and in-vivo tissues.