Identification of a Complex Drive Chain Based on Local Linear Model Tree

Abstract

This paper uses the local linear model tree (LOLIMOT) method for modeling the angular velocity of a complex nonlinear system called Gamma-log. The drive chain of the Gamma-log contains nonlinear parts such as an AC servo drive or a worm gear drive. The drive chain is modeled with LOLIMOT algorithm. An experiment was conducted to collect data from the original system and to simulate the kinematics of the track. The best model was selected from ARX and FIR models using a correlation coefficient based performance index.