Remote-Center-of-Motion Generation for Hands-on Robotic Surgery: A Constrained Optimization and Neural Network Approach

In robot-assisted minimally invasive surgery, the robot needs to hold a long instrument that is inserted through an incision in the patient's body, and the motion of the instrument should meet remote-center-of-motion (RCM) constraints. This paper presents a quadratic programming (QP) based control scheme for hands-on robotic surgery. The proposed scheme enables simultaneous generation of the RCM considering the insertion of tool shaft, robot passivity under admittance control, and joint-limit avoidance. Both simulative and experimental validations have been successfully conducted. A linear-variational-inequality based primal-dual neural network (LVI-PDNN) is applied to solve the QP problem, validating that the proposed QP-based scheme can meet RCM constraints when the robot is guided by an external force applied by a human operator while simultaneously considering the insertion limit and joint constraints.