Study On Robot-Assisted System For Percutaneous Surgery Based On 3D Ultrasound Images

Due to minimal trauma and pain, few complications and fast recovery compared with conventional surgery, percutaneous surgery is more and more welcomed by doctors and patients especially in the fields of biopsy and interventional therapy, and has been widely used in treatment of tumor resection as non-surgical effective method. Since ultrasound is real-time, non-ionizing, economical and convenient, it has been widely used to guide clinical percutaneous surgery. However, the ultrasound-guided percutaneous surgery has some deficiencies such as limited needle insertion angle, low accuracy and stability, and over-reliance on medical experience. Therefore, the technology of image-guided robot-assisted percutaneous surgery has become one of international hot research issues in recent years. This dissertation, supported by Key Project of Heilongjiang Natural Science Foundation, performs researches on key technologies of robot-assisted system for percutaneous surgery guided by freehand 3D ultrasound, which mainly includes system construction, 3D reconstruction of freehand ultrasound images, motion planning and control of percutaneous robot, motion parameters calculation of percutaneous robot, and operation monitoring.For clinical applications of ultrasound-guided percutaneous surgery, the robot-assisted system for percutaneous surgery guided by freehand 3D ultrasound images is constructed in this dissertation, which consists of 6-DOF percutaneous robot subsystem, freehand 3D ultrasound subsystem and electromagnetic tracking subsystem. According to the planning motion steps and calculated motion parameters, the robot can perform percutaneous surgery instead of human operation, which overcomes hand tremble, improves surgical accuracy and stability, and reduces labor intensity of doctors. Freehand 3D ultrasound can create 3D geometric model of lesions to help doctors plan surgery more accurately, which overcomes the limitations of 2D ultrasound, and reduces dependence on medical experience and impact from human factors. Electromagnetic tracking system replaces the clinical guiding device, which can both solve the problem of limited needle insertion angle, and realize spatial real-time spatial tracking and monitoring of surgical instruments.The key technologies for 3D reconstruction of freehand ultrasound images include probe calibration, calculation of voxel intensity and visualization of 3D regular volume grid. Via analyzing the probe calibration problem of N-shaped model, a new probe calibration method based on improved N-shaped model and non-iterative least squares solution is presented. The improved N-shaped model can linearize and simplify calibration equation, and avoid local convergence and inappropriate initialization of iterative value. Based on the analysis of PNN algorithm performance, the PNN-Maximum algorithm is proposed to calculate voxels intensity, which can save time of searching vacant voxels and avoid error caused by interpolation. Based on the analysis of characteristics of common algorithms for 3D image visualization, Ray-Casting algorithm in VTK is used to realize volume rendering of 3D regular volume grid. The experiment validates the effectiveness of freehand 3D ultrasound in this dissertation.According to clinical percutaneous surgery procedures, surgical safety and medical operation habits, motion steps and motion function of percutaneous robot is planned, in which robotic motion function is achieved by appropriate control methods. Robot motion is divided into four steps according to following sequence: arm translation, wrist rotation, needle insertion, and needle withdrawal. Taking into account the clinical practices and surgical safety, the force-driven drag control method for percutaneous robot is presented, which converts drag force to pulse signal of motors, so doctors can drag the robot arm to move the needle. Experimental result shows the robotic drag control has good compliance. Based on the analysis of needle orientation adjustment methods, the needle tip displacement compensation algorithm based on Cartesian reverse movement is proposed, which uses reverse movement of robot arm to compensate real-time needle tip displacement caused by robotic wrist rotation so as to maintain the needle tip fixed while robot wrist is rotating. Experimental results show the needle tip displacement is less than 1mm, which verifies the effectiveness of this algorithm. According to the experience knowledge of clinical manual operation, the fuzzy control method of robotic needle insertion speed based on force and position feedback is presented. In this method, needle insertion depth and needle insertion force are two input of fuzzy controller, and needle insertion speed is output of fuzzy controller. Experimental results show the needle insertion control strategy can effectively improve the safety of robotic needle insertion and reduce impact of soft tissue deformation on robot-assisted percutaneous surgery.In order to map the needle path in 3D image space to robot space and calculate robot motion parameters, spatial mapping of robot-assisted system for percutaneous surgery is established with the electromagnetic transmitter as reference coordinate system. Then, the needle can be monitored in 3D image space by the technology of needle calibration. To verify effectiveness of the robot-assisted system for percutaneous surgery guided by freehand 3D ultrasound, a series of experiments are made to test needle positioning accuracy. Experimental results show that overall positioning error of robot-assisted system for percutaneous surgery is within 5mm, which can satisfy the medical requirement for abdominal percutaneous surgery. Finally, the positioning error of the robot-assisted system for percutaneous surgery is analyzed, which provides the references for design and improvement of this kind of system.