Nikola Živković

Lola Institut · Robotics

A robot dynamic model is time variable, highly non-linear and characterized by coupling effects among the robot joints. Consequently, a derivation and implementation of a robot dynamic model, which is used for purposes of control, simulation, and mechanical design, often represents a challenging task. The last couple of decades saw a great amount o...

This paper presents a method for the implementation of a robot forward kinematics algorithm that complies with the Denavit-Hartenberg (DH) convention in Robot Operating System (ROS). The integration of the algorithm in ROS is based on the representation of DH parameters in dual quaternion space. The main motivation for the presented research is to...

In this paper an advanced iterative learning control algorithm for rehabilitation exoskeletons is proposed. A simplified biomechanical model is used as the control object to verify control algorithm feasibility. The control design is proposed as two level controller consisting of inner and outer loop. In the inner loop the feedback linearization is...

The paper designs an appropriate iterative learning control (ILC) algorithm based on the trajectory characteristics of upper exosk el eton robotic system. The procedure of mathematical modelling of an exoskeleton system for rehabilitation is given and synthesis of a control law with two loops. First (inner) loop represents exact linearization of a...