Dynamic Analysis and Controller Design for the Ballbot

Abstract

A ballbot is a dynamically stable mobile robot that shows the capability of omnidirectionality, agility, and maneuverability on a floor. This paper presents the design prototype and control scheme of a ballbot. By assuming that the effected noise is white gaussian noise, a Kalman estimator is applied to estimate the internal state of the system. To improve the trajectory tracking performance of the ballbot, a Linear Quadratic Tracking (LQT) controller is designed to balance and transfer the ballbot system taking into account the presence of the noise in system, and the numerical simulation results imply the correctness of the system modeling and effectiveness of LQT control design for a MIMO ballbot system.