Autonomous Applied Robotics: Ultrasound-Based Robot-Assisted Needle Insertion System Concept and Development

Takács, Bence and Haidegger, Tamás (2020) Autonomous Applied Robotics: Ultrasound-Based Robot-Assisted Needle Insertion System Concept and Development. In: IEEE 15th International Conference of System of Systems Engineering (SoSE 2020), Piscataway (NJ), Amerikai Egyesült Államok.

Preview

Text sose2020_102_Takacs.pdf.pdf Download (5MB) | Preview

Abstract

Ultrasound (US) is a popular imaging modality for image-guided minimally invasive surgery (MIS), enabling the faster and more reliable execution of numerous procedures, such as biopsy, electrode placement and vessel cannulation. Blood vessel cannulation is a common, routine intervention, e.g., for blood oxygen level testing. Yet, in particular cases, when the vessel is located deep or veins less stable (with the loss of subcutaneous tissue), it is hard to complete it without US assistance. In this paper, we present a solution for US-guided, robot-assisted needle insertion for vein cannulation. We developed an image-guided system to aid needle insertion via active targeting and anatomy-relevant positioning, together with safeguarding features, such as a kinematically enforced Remote Center of Motion (RCM) mechanism. The proposed system comprises a portable US transducer mounted on a KUKA iiwa collaborative robot, a custom designed needle insertion mechanism with adjacent controllers. The US and needle insertion mechanism are attached to the robot through a 3D printed custom designed mounting part with integrated force sensor. The robot arm is responsible for moving the needle to target position with impedance control. The needle insertion mechanism allows the manipulation of the needle along 3 axes. The mechanism was designed for near-surface vein cannulation with an RCM kinematic structure to avoid damage to the vein. The developed system was tested with different types of gelatin phantoms. Vein deformation and tissue motion was examined during US imaging. The control loop of our system is supplemented with vein deformation tissue model and US-based visual servoing.

Actions (login required)

Edit Item