Dynamic modeling and adaptive neural network control of a class of 3-dof tendon-driven minimally invasive instruments

Abstract

To get compact instruments in robot-assisted minimally invasive surgery (MIS), each servo motor was installed the base and motor torque was transmitted to each joint through a tendon-pulley system. Trajectory tracking control is very important in MIS. In this paper, dynamic structure and equation of motion including the effect of rotor inertia for a class of 3-dof tendon-driven minimally invasive instruments were established. An adaptive controller based on model block approximation RBF neural network was designed for a class of 3-dof tendon-driven minimally invasive instruments. In addition, the neural network modeling errors can be easily suppressed by incorporating robust control. Trajectory tracking control numerical simulation was carried out. The simulations results show that validity and effectiveness of the derived model and designed adaptive neural network.