Ultrasound Image Guided Robotic Puncture System With Needle Tip Detection and Deviation Correction Features

Abstract

Robotic needle insertion with ultrasound image guidance can help improve the accuracy of liver cancer biopsy. Most robotic puncture systems with ultrasound guidance utilize dual-arm configurations, which are less efficient in terms of space utilization. To address this, we have developed a compact ultrasound image guided needle puncture mechanism. It consists of an ultrasound probe to provide image feedback and a serial-parallel mechanism together with a range extender to drive the puncture needle. To achieve accurate puncture in the dynamic environment of tissue deformation and needle bending, a needle tip detection algorithm in ultrasound images with YOLO V8 network is proposed, together with needle axis fitting and Kalman Filter to furtherly improve the accuracy. With image feedback, a closed-loop deviation correction controller with fuzzy control algorithm is developed and tested on the robotic puncture system, achieving 0.83mm accuracy, which can meet the clinical requirement.