Related Experiment Video
Updated: May 8, 2025

Design and Implementation of a Bespoke Robotic Manipulator for Extra-corporeal Ultrasound
Published on: January 7, 2019
Development of a Low-Profile, Piezoelectric Robot for MR-Guided Abdominal Needle Interventions
Samuel Wilcox1,2, Saikat Sengupta3, Chuan Huang4
1Institute of Robotics and Intelligent Machines, Georgia Institute of Technology, Atlanta, GA, 30332, USA.
Purpose:
Minimally invasive needle-based interventions are commonly used in cancer diagnosis and treatment, including procedures, such as biopsy, brachytherapy, and microwave ablation. Although MR-guided needle placement offers several distinct advantages, such as high-resolution target visualization and accurate device tracking, one of the primary limitations that affect its widespread adoption is the ergonomic constraints of the closed-bore MRI environment, requiring the patients to be frequently moved in and out to perform the needle-based procedures. This paper introduces a low-profile, body-mounted, MR-guided robot designed to address this limitation by streamlining the operation workflow and enabling accurate needle placement within the MRI scanner.
Methods:
The robot employs piezoelectric linear actuators and stacked Cartesian XY stages to precisely control the position and orientation of a needle guide. A kinematic model and control framework was developed to facilitate accurate targeting. Additionally, clinical workflow for the liver interventions was developed to demonstrate the robot's capability to replicate existing procedures. The proposed system was validated in benchtop environment and 3T MRI scanner to quantify the system performance.
Results:
Experimental validations conducted in free space demonstrated a position accuracy of 2.38 ± 0.94 mm and orientation error of 1.40 ± 2.89°. Additional tests to confirm MR-conditionality and MR-guided phantom placements were carried out to assess the system's performance and safety in MRI suite, yielding a position error of 2.01 ± 0.77 mm and an orientation error of 1.57 ± 1.31°.
Conclusion:
The presented robot shows exceptional compatibility with a wide range of patients and bore sizes while maintaining clinically significant accuracy. Future work will focus on the validations in dynamic liver environments.

