Kalman Filter-based compensation of disturbance forces and torques for AGVs

Abstract

Controller design for autonomous 4-wheeled ground vehicles is performed with differential flatness theory. Using a 3-DOF nonlinear model of the vehicle's dynamics and through the application of differential flatness theory an equivalent model in linear canonical (Brunovksy) form is obtained. The processing of velocity measurements (provided by a small number of on-board sensors) through a Kalman Filter which has been redesigned in the form of a disturbance observer results in accurate identification of external disturbances affecting the vehicle's dynamic model. By including in the vehicle's controller an additional term that compensates for the estimated disturbance forces, the vehicle's motion characteristics remain unchanged.