Intelligent Control of a Master-Slave based Robotic Surgical System

A master-slave (M-S) based free environment (in-air) on-line trajectory control of a redundant surgical robot is presented in this work which can provide solution to precise positioning of the surgical robot tip that can be deployed for the internal organ inspection, radiosurgery of the gastrointestinal tumor, improvement in the precise placement of radiation source during brachytherapy, etc. This work deals with the online in-air 3D trajectory tracking control of a cable-driven 5-degrees of freedom (DOF) redundant slave robot deployed in the M-S based robotic surgical system presented towards robotic natural orifice transluminal endoscopic surgery (NOTES) purpose. A novel intelligent control technique involving an uncomplicated supervised radial basis function network (RBFN) based mapping is proposed to learn the inverse kinematics (IK) relationship of the redundant slave robot for tracking the desired 3D trajectory, which is given by the master robot under in-air environmental conditions. The proposed method eliminates the reckoning of Jacobian pseudo-inverse of the slave robot in real-time. To resolve the redundancy of the 5-DOF slave robot, a generalised approach in the weight update law during the learning phase of the inverse kinematic control is proposed. The proposed control algorithm’s trajectory tracking performance is compared with the back propagation network (BPN) based approach and the gold standard closed-loop inverse kinematic (CLIK) redundant control algorithm involving instantaneous maximization of manipulability measure through both simulation and experimental studies. Root mean square error (RMSE) of 0.572 mm (simulation) and 0.600 mm (experiment) are obtained for the RBFN based approach as compared to that of the BPN (simulation: 0.910 mm and experiment: 1.500 mm) and CLIK (simulation: 0.607 mm and experiment: 0.900 mm) algorithms are reflecting its best performance towards trajectory tracking tasks. Thus, the novelty of the current study lies on the development of a simple, uncomplicated intelligent neural network controller overperforming in 3D trajectory tracking tasks by the standard kinematic control techniques such as CLIK and other model based techniques. The experimental setup for M-S based surgical system for robotic-NOTES is developed to validate the intelligent control technique.