| Needle insertion is one of the common minimally invasive surgical procedures,such as biopsy,local anesthesia,brachytherapy and interventional radiologies.For most cases of diagnosis,treat-ments and researches,the needle tip is required to place to the specific target,which localizes the soft organs.The insertion errors caused by the uncertainties of target motions and needle trajec-tories limits the clinical application of needle insertion.The goal of the dissertation is to let flexible needles hit the targets accurately.The needle-tissue interactive mechanism is investigated,and a finite element model(FEM)for soft tissue de-formation is presented with the modified local constraint method.The 2-dimensional simula-tion is implemented to predict soft tissue deformation with MATLAB program package,and it is showed that the algorithm is valid from the experimental results.The computer experiment anal-ysis method is employed to estimate the deformation of soft tissue,and the Kriging-based model is created to predict the real-time information of targets and obstacles.Compared to the results of finite element model,the Kriging-based model could overcome the time-consuming drawbacks.The kinematics of flexible needle,which describes the relationship between the base and the tip,is modeled with the Denevit-Hartenberg(D-H)method.The equation of inverse kinematics is constructed with the adjusting angles,and the experiments show the effectiveness of the forward and inverse kinematics model.The general insertion error is defined with the concept of artificial potential field,which considers the motions of targets and obstacles inside soft tissue.The path planning problem in the dynamic environment is converted into the optimization of general inser-tion error.The algorithm for dynamic path planning is presented combined with the forward and inverse kinematics,and the experiments were carried out for validating the feasibility of proposed path planning algorithm.This dissertation consists of the following six chapters.In Chapter 1,the background and state-of-the-art of computer aided needle insertion into soft tissue are introduced in detail,and the classifications and exhibitions of insertion errors are concluded.A comprehensive survey,including the needle-tissue interactive mechanism,flexible needle steering technologies and path planning algorithms,is given.The challenging problems of soft tissue deformation and path planning are pointed out.The objectives and main contents are summarized finally.In Chapter 2,the soft tissue is modeled with the linear elastic finite element method,and the flexible needle is modeled as a cantilever beam.The coupling model of flexible needle and soft tissue is constructed with the modified local constraint method,with which the forces acting on needle nodes could be transmitted to the tissue nodes.The procedure of needle insertion into soft tissue is simulated with MATLAB software.The experiment is designed with the simulation conditions,and the displacements of tissue nodes are recorded by the markers inside the PVA phantoms.Three types of tissue nodes,including the nodes constrained by the needle,close to the boundary and far away from the needle,are illustrated the effectiveness of the proposed simulation model.The average error between the simulation and experimental results is less than 0.50mm.The proposed coupling algorithm could also be applied to the nonlinear and inhomogeneous soft tissue materials.In Chapter 3.the computer experiment analysis and metamodel are employed to estimate the motions of targets and obstacles inside the tissue.The basic principle of Kriging metamodel and Latin hypercube sampling method are detailed.The Kriging-based estimation model for the maximum value of tissue deformation is constructed with 11 variables,which involves the insertion angle,the material parameters of needle and tissue.The effects of correlation functions on the predictability and parameter sensitivity are analyzed.The real-time deformation of soft tissue is predicted with the functional response Kriging-based model.The real-time Kriging model is analyzed for adapting the changes of parameters and time indices.Compared with the FEM results,the relative residual error of real-time prediction is less than 35%and the Kriging-based model could reflect the mechanism of soft tissue deformation.In Chapter 4,the forward and inverse kinematic models of flexible needle are presented with the robotic theory based on the analysis of needle deflection.According to the quasi-static thinking,the procedure of needle insertion is decomposed into n sub-procedures and the needle is divided into n needle segments.The motion of needle tip consists of two rotations and one translation for each sub-procedure.The relationship between the base and the tip is described as the homogeneous transformation with D-H method.The setup is constructed for validating the kinematic model.The tip offset error between kinematics prediction and experimental result is less than 0.80mm.and the targeting error is less than 0.78mm.The kinematic model could describe the needle deflection inside soft tissue.In Chapter 5,the path planning algorithms are proposed for the static environments with and without obstacles based on the existence analysis of inverse kinematical solutions.The general insertion error is defined as the combination of target hitting and obstacle avoidance with the con-cepts of Euclidean distance and artificial potential field.The path planning problem in the dynamic environment is converted into an optimization problem of general insertion error.Combined with the Kriging-based tissue prediction model,the dynamic path planning algorithm is presented for the environment with moved obstacles and targets.The experimental results show that a feasible trajectory could be obtained with the static and dynamic path planning algorithms,which consider both the steerability of flexible needle and the feasibility of insertion trajectory.In Chapter 6,the research work of this dissertation is summaried,and some future directions of needle insertion into soft tissue are discussed. |
