Research On The Reliability Of MRI Surgical Robot System

This paper is focused on the Magnetic Resonance Imaging (MRI) guided needle in-sertion robot for prostate brachytherapy, aiming to improve its reliability both in structureand kinematics. Accounting that available data information of surgical robot is alwaysinsufcient, this paper makes modification and improvement to the design of robot by com-bining methods of theoretical calculation, software simulation and experimental validation.The main contents contain a new robust optimization to the dimension of robot based ondensity-convex reliability model, and a new kinematic calibration method and control com-pensation strategy based on Screw Theory. The main research contents and results achievedare described as follows:Firstly, the density-convex model is first proposed on account of the reality that avail-able data information is always insufcient. By using Finite Element Method and Artifi-cial Neural Network (FEM-ANN) method, the explicit functions of mechanical responseare achieved efectively. Then reliability functions of structure are constructed based ondensity-convex model. The optimization of robot dimensions is carried out by taking robustindex as the ultimate objective function and reliability of structure as constraint conditions.Secondly, a novel calibration process based on Screw Theory is proposed where thenonlinear efects of joint compliance are considered. Based on screw theory, the kinematicerror model is established for the relationship between error of controller and the deviationof the measured pose of the end-efector, synthesizing actuator error, geometry error, andjoint error. Therefore the accuracy of robot can be improved by adjusting parameters ofcontroller real-timely based on the control strategy for compensation.Finally, strain tests and position measurements of needle-tip are conducted on anexperimental platform including MR(Magnetic Resonance)-compatible robot and controlsystem. Strain tests at the weak links are conducted to validate the structural reliabilityand structural robust by the comparison between theoretical and actual strain. Comparativemeasurements of needle-tip trajectories before and after compensation validate the kine-matic reliability and efectiveness of proposed calibration and compensation approaches.The research results of this paper have important theoretical significance and value forpractical application in the design of MRI-guided surgical robot. The results described inthe paper also lay a sound foundation for the future work to improve the reliability of robotsystem in magnetic field.