Research On Motion Planning And Control Of Venipuncture Robot

As a common clinical treatment method,venipuncture can only be completed by nurses at present.However,there are a lot of patients in hospital,and nurses are in short supply.The success rate of venipuncture in elderly,children and obese patients is low.In order to meet the needs of people for better medical services,an intelligent venipuncture robot is urgently needed.At present,there are some problems in the venipuncture robot,such as low intelligence,difficult to obtain the position and pose information of the target puncture vessel.In this paper,based on a six axis venipuncture robot platform,a binocular vision measurement system based on near-infrared camera is developed.The theoretical and Experimental Research on the acquisition of target vessel characteristics and spatial pose information,and the path planning of the robot toward the target vessel are carried out.The main research contents and results are as follows:1.Blood vessel feature extraction and target recognition.Firstly,based on the binocular near-infrared vision system,the left and right hand back blood vessel images are collected.Secondly,the camera calibration model is used to correct the image,and a series of preprocessing methods such as ROI region extraction,image denoising,enhancement,segmentation,mathemati-cal morphology are’ used to extract veins,and the relatively smooth and complete veins are obtained by CLAHE enhancement and adaptive threshold method.Then,based on the eight neighborhood features of image pixels,the vascular features and geometric information,including vessel centerline,branch,edge contour,diameter and curvature,were calculated respectively,and the effective vascular feature information was extracted based on Zhang thinning algorithm and connected domain labeling method.Finally,according to clinical venous puncture experience,the target puncture was obtained by comparing the diameter and curvature of vessels Vascular area and location.2.Three dimensional reconstruction of the target puncture vessel and estimation of the position and posture of the puncture point.Firstly,according to the principle and method of stereo vision measurement and camera calibration,the internal and external parameters of the binocular near-infrared camera are calibrated.Secondly,through the analysis and comparison of various vascular image stereo matching methods,a point matching algorithm based on epipolar and similarity constraint is adopted.The algorithm greatly reduces the complexity of vascular matching and matches the blood vessels.Then,through the limit constraints and similarity constraints between the left and right image vascular centerline points,the correct matching points of vascular centerline are selected,and the 3D reconstruction of vascular centerline is carried out by using binocular triangulation principle.According to the estimated two-dimensi onal half diameter of blood vessel,the vessel wall is fitted with a spatial circle with the vascular centerline as the axis.After that,the position and posture of the puncture point were estimated according to the angle of the clinical needle.3.Research on robot kinematics modeling and path planning.Firstly,the joint coordinate system of the connecting rod robot is established according to the D-H rule,and the forward and inverse kinematics equations of the robot are solved.Then,the physical motion model of the robot is established by SolidWorks and ADAMS software to create a platform environment for robot simulation,and the correctness of kinematics is verified by simulation.Finally,the process of the needle towards the target blood vessel puncture point is conducted by using the quintic polynomial method Path planning.The simulation results show that the motion of the robot is stable,and the displacement,velocity and acceleration curves of each joint are consistent with the planning results,which verifies the feasibility and effectiveness of the method.4.Development and experiment of venipuncture robot system.Firstly,the experimental platform of venipuncture robot is developed,and the hardware control system and software control platform based on the fixed height controller are designed and built.Then,the near-infrared hand eye camera is calibrated by Tsai Lenz method,and the hand eye camera calibration matrix is obtained,and the position and posture of venipuncture point in camera coordinate system are converted to robot coordinate system.Finally,the machine is completed In the experiment of human vein puncture path planning,the robot is controlled to reach the position of human blood vessel puncture point.At the same time,the robot calibration error is mainly from the vision system of the human body,blood vessels and so on.The experimental results show that the robot can accurately reach the given puncture point position and preliminarily realize the semi-automatic venous blood collection.