Research On Needle-Soft Tissue Interaction,Path Planning And Sensor Integration Of Robotically Assisted Needle Placement

The mainstream clinical procedures for diagnosing and treating cancer and other pathological tissues include percutaneous interventions,among which clinicians insert surgical needles to target regions inside patients either to take tissue samples for diagnosis?biopsy?,or to place radioactive seeds for treatment?brachytherapy?.Needle placement is a typical Minimally Invasive Surgery?MIS?,and it is widely adopted for its unotable trauma,fast recovery and remarkable flexibility.However,the target region?for example a tumor?is inside the patient's body,which makes the whole procedure non-transparent and the operation requires high-quality intraoperational medical images.On one hand,Computed Tomography?CT?,Magnetic Resonance Imaging?MRI?and Ultrasound Imaging?US?all have certain limitations in speaking of environment and material requirement,and are not satisfying in the terms of real-time performance and resolution.On the other hand,the commonly used bevel-tip needles bend dramatically during insertions,which makes preoperatedly designed straight paths inapplicable and increases difficulty for bare-hand operations.In the past decades,medical robotic techniques have significantly enhanced surgeons'capability,with respect of stability,safety and controllability.Therefore,in this thesis we propose theories and techniques for robot-assisted needle placement.The main scopes of the thesis include needle deformation mechanism,path planning,sensing techniques and control approaches.Two towards-clinical robot system prototypes are developed to validate the proposed methods and to lay the foundation for future researches.The main contents of this thesis are as follows:Flexible needle-soft tissue interaction mechanism.Force distribution on stiff and flexible needles are analyzed and based on which,needle tip motion prediction approaches are studied.This part is about the kinematics of the needle's targeting problem,and accurate prediction of the needle tip position is essential for path planning.Path planning and needle control methods.Multi-objective optimized path planning methods are developed for both stiff and flexible needles,especially a novel path searching approach based on the recollection probe algorithm.Several insertion experiments are carried out to validate different interactive control methods.Basically,this part is about the inverse kinematics of the needle's targeting problem.Needle shape sensing based on integrated FBG sensors.To alleviate the problem aroused by ineffective intraoperational images,a Fiber Bragg Gratings?FBG?integrated flexible needle is devised,fabricated and calibrated to obtain accurate real-time shape of the needle.A 4^th order shape reconstruction method is proposed and validated by comparing with other approaches in literature.Interactive control based on real-time information feedback.An interactive control approach is proposed for targeting in multi-layer soft tissues,in which FBG information is used to detect model parameters for unknow soft tissues and to predict and update needle tip paths.Targeting experiments in two-layer phantoms demonstrate the effectiveness of the approach.Moreover,experiments on towards-clinical needle placement robot system prototypes are carried out as tentative efforts for future clinically-practical robot-assisted needle placement.