Objective To explore the accuracy of the multiplication coefficient method and the moment synthesis method for determining the spatial position of the center of gravity(CoG)of the human body based on machine vision.Methods The mechanical measurement platform was built,and the three-dimensional(3D)human body CoG measurement method under static and dynamic conditions were designed to calculate the space coordinates of the CoG.Through experiments,the calculation accuracy of the multiplication coefficient and moment synthesis method were studied and analyzed.Results In the static experiments,the calculation results of the torque synthesis method were more accurate than those of the multiplication coefficient method for each dimension.The errors in the 3D CoG of the human body in the X,Y,and Z directions calculated using the torque synthesis method were 3.9%,4.1%,and 8.5%,respectively.In the dynamic experiment,the average and relative errors of the torque synthesis method in the X or Y direction were lower than those of the multiplication-coefficient method.When the action decomposition method was used to analyze the height direction of the CoG along the Z axis,the final rendering effect of the torque synthesis method improved.Conclusions The accuracy of the 3D CoG calculated by the moment synthesis method was relatively high,and was closer to the measurement data of the mechanical measurement platform.The 3D CoG calculated using the moment synthesis method can replace the mechanical measurement platform and can be used in subsequent studies.

Objective To propose a new multi-joint series venipuncture system, explore the mechanics and kinematics-based related control problems involved in needle insertion and needle picking during the puncture process, and verify feasibility of this system. Methods A puncture manipulator was built, and needle displacement control algorithm was proposed by combing with the puncture mechanics model. The the forward kinematics was calculated by using DH method, so as to obtain the tip coordinates. Then the inverse kinematics was calculated by using the geometric method. The forward and inverse processes were closely connected. The position error of the end coordinates before and after needle picking was compared by using the method of kinematics positive solution-inverse solution-re-positive solution. Finally, experimental verification and simulation were conducted by combining with the physical object. Results Through simulation and experiments, accuracy of the theoretical model was verified. The needle insertion algorithm could be used to achieve success with only one needle insertion, which provided theoretical basis for the control of robot arm. The position error before and after needle picking could be controlled within 1 mm from the end trajectory. The end needle tip of robot arm was almost kept fixed during the needle picking process. Therefore, this needle picking scheme was feasible and could basically verify that the needle picking action of robot arm met the accuracy and safety requirements. Conclusions The venipuncture manipulator truly simulates the needle insertion and needle picking action during the puncture process, and can safely and accurately realize the needle insertion and needle picking action with needle tip as the fixed point, indicating that it has certain clinical value.