Comparisons of RCM Generation Algorithms for Vision-Controlled Robotic Endoscope

Abstract

In minimally invasive surgery (MIS), a surgical endoscope is an essential instrument that provides visualization for the surgeon. One principal characteristic of surgical instruments is that remote center of motion (RCM) must be respected. To meet such a practical requirement, many physical RCM mechanisms and software-based RCM generation algorithms have been proposed. As compared with physical RCM mechanisms, RCM generation algorithms possess more flexibility due to the fact that the RCM point can be adjusted if required. This paper conducts comparisons of four typical RCM generation algorithms applied to a vision-controlled robotic endoscope under joint constraints. Kinematic models of the robotic endoscope and the four RCM generation algorithms are first briefly introduced. Then, a unified control formulation based on quadratic programming (QP) is constructed to incorporate kinematic, RCM, and physical constraints of the robotic endoscope. Based on the unified control scheme, comparative simulations and experiments are performed. The advantages and disadvantages of the four typical RCM generation algorithms are analyzed and discussed. When performing a same peg transfer task in the simulations, the RCM errors synthesized by RCM generation algorithms designed using a plane equation and an insertion equation are smaller. In the physical experiments, there are few differences in the RCM errors. Nevertheless, it is revealed that the joint velocities corresponding to the RCM generation algorithm based on a plane equation are the smallest, which means that the joint angles change more gently and it can be more friendly to MIS.