Dynamic modeling and control of a 4 DOF robotic finger using adaptive-robust and adaptive-neural controllers

In this research, first, kinematic and dynamic equations of a 4-DOF 3-link robotic finger are derived using Denavit-Hartenberg convention and Lagrange’s formulation. To model the muscles, several springs and dampers are placed between the finger links. Then, two advanced controllers, namely adaptive-robust and adaptive-neural, which can control the robotic finger in presence of parametric uncertainty, are applied to the dynamic model of the system in order to track the desired trajectory of tapping. The simulation of the dynamic system is performed in presence of 10% uncertainty in the parameters of the system and the results are obtained when applying the two controllers separately on the robotic finger dynamic model. By comparing the simulation results of the tracking errors, it is observed that both controllers perform decently; however, the adaptive-neural controller has a better performance.