Modelling and optimization of robotic manipulator mechanism for computed tomography guided medical procedure

Although industrial robots are common, a higher degree of manipulability might be required to expand the applications of manipulators in the field of medicine. Modifying the mechanical design of a robot as per the workspace can be perceived as an optimization problem. Hence, a novel spatial manipulator is designed for a diagnostic apparatus using different optimization algorithms. Standard Genetic Algorithm (SGA) and GA (Genetic Algorithm) with hybrid functions like pattern search (PS) and fmincon are proposed to optimize the link lengths of a 3degrees of freedom (DOF), 6-DOF, and novel 9-DOF hybrid redundant manipulator. A 9-DOF robot is designed to manipulate a needle in CT machine environment. The fitness function for all the manipulators is formulated using forward kinematic equations according to their workspace. Limits and constraints of each link are decided beforehand. A comparative study between all the hybrid GA functions is performed. MATLAB is used to solve and train the proposed GA method for optimizing the link lengths. Results show that GA with PS provide better-optimized link lengths for a 3-DOF and 9-DOF manipulator while fmincon is well suited for a 6-DOF robot manipulator. Workspace and dead zone analysis is also performed using the optimized link lengths obtained.