Study On Information Collection And Control Technology Of Robot For Liver Cancer Puncture Ablation

Despite of navigation systems and surgical robots have applied in orthopedics,neurosurgery,otolaryngology,and other fields,doctors are still locating tumors by eyes and puncturing by hands in liver tumor ablation surgery.The reason why navigation systems and surgical robots are not applied in liver tumor ablation is that they are flawed.The core component of navigation systems are optical tracking systems(OTS),which have the problem of generating ghost markers including false recognition and false corresponding,and ghost markers will affect the operation process.At the same time,commercial neurosurgery robots are not able to ensure the accuracy of liver tumor puncture because of the liver deformation and displacement in respiratory movement.In order to improve the fluency and safety of the optical guided surgery,it is necessary to solve the problem of generating ghost markers in OTS.Besides,it is necessary to study the robot pose calculation method and robot control algorithm to compensate respiratory movement and to improve the accuracy of puncture.This article completes the corresponding research for the above problems:(1)The methods of removing ghost markers are studied.The methods of image segmentation and feature screening are used to remove false recognition in object detection.Then 3D coordinates are estimated by using the camera pinhole model which can remove the false corresponding.These methods improve the robustness of OTSs and reduce the calculation load of the subsequent navigation process.It is proved that the performances of our method in removing ghost points are better than commercial OTSs' by the experiments.(2)The optimal target pose calculation method is studied.We use the constraint relationship between the planned puncture path and robot to find the optimal pose of the robot,and make the robot drive the puncture needle along the planned path.Then retain the unique target joint angle in the inverse kinematics to avoid the joint flipping when moving.The average error of the robot puncture in puncturing rigid phantom experiments is within 1mm,which proves the accuracy of the target pose calculation method.(3)The control method of the puncture robot is studied.A compound control method combining PID and multiple feedback is used to construct a non-linear control system,which enables the robot to track moving target in real time.Combined with the optimal time puncture method proposed by our team,the robot can insert the needle in the optimal time that matches best the patient and CT images.Finally,the experiments of manual puncture and robot puncture by respiratory phantom were carried out.The average of manual puncture error was 10.3mm,and the average of robot puncture error was 4.9mm.