Research On Optimization And Speech Control Of Auxiliary Manipulator For Hysterectomy

Hysterectomy,as a kind of minimally invasive surgery,improves the problems of large wound and obvious scar in traditional surgery,and has the advantages of small trauma,fast wound healing and short recovery time after surgery.With the development of robotic technology in the medical field,it has become a trend to use robot assistants doctors to complete surgery.This dissertation takes hysterectomy as the background,realizes the remote center of motion(RCM)in the surgery process with the help of KINOVA manipulator,focuses on the remote center location of the manipulator in the surgery process,and realizes the intelligent speech control mode based on the neural network model.Doctor controls the manipulator by speech command in the surgery,the surgery efficiency is improved.The successful implementation of the RCM during hysterectomy is the basis for the completion of the surgery.This dissertation will use KINOVA manipulator to carry out surgery.The manipulator has six degrees of freedom and can be well programmed to achieve RCM.The location of remote center in the surgery process has a certain influence on the flexibility and manipulability of the manipulator.How to select the optimal remote center location is the key to the surgery.In this dissertation,the reference coordinate systems of the manipulator are established,and the D-H parameters are obtained.Then the forward and inverse kinematics solutions of the manipulator are obtained.By further analyzing the augmented forward kinematics of the manipulator,an augmented Jacobian matrix with remote center constraint is obtained.An optimization index of global isotropy is established to search the optimal remote center location in the workspace.The key to successful surgery is to complete the interaction between doctor and manipulator during surgery.Speech control as a convenient human-machine interaction technology,this dissertation analyses the traditional speech recognition methods,according to the requirements of hysterectomy,propose a speech recognition technology based on convolutional neural network.Speech signals need to be preprocessed before speech recognition.In order to grasp the features of speech better,MFCC feature parameters are extracted.Mel spectrum is drawn as the input of convolution neural network to give full play to the advantages of convolution neural network.The control interface of human-machine interaction is written to facilitate speech input,recognition display and experiment simulation.The simulation control platform of the manipulator is built under ROS,and the RCM of the manipulator is simulated with MoveIt! function package,and we also verify the accuracy of the speech recognition system.The experiment results show that the system can recognize the speech commands correctly and drive the machine to complete the corresponding motion in real time.The exploration and research of remote center localization and speech control of manipulator for minimally invasive surgery have reference significance for other people in this field.