Compensation of friction in robotic arms and slide tables

Authors:	Zabinski Tomasz, Rzeszow University of Technology, Poland Turnau Andrzej, University of Science and Technology Kraków, Poland
Topic:	4.2 Mechatronic Systems
Session:	Mechatronic Applications
Keywords:	Direct-Drive Robots, Friction, Compensation, PD-Controllers, Neural control, Positioning systems

Abstract

The paper presents experimental results of friction parameter identification for two different systems: rotational arms of a direct-drive spherical manipulator and linear slide tables. Parameters of the static part of the LuGre friction model are estimated. Several experiments are reported in which selected friction compensation methods are employed for tracking at low velocities. Two compensators working in real time are constructed and verified: the PD-like lead, to avoid stick-slip motion and the on-line trained neural network with a special structure, to avoid errors caused by friction and varying dynamics of robotic actuators.