Skip to content

Decoupled Dynamic Control for Pitch and Roll Axes of the Unicycle Robot

Document Type
Publications · Journal Article
Registered DB
SCIEScopus
Journal Ranking
JCR Q1Scopus Q1
Journal
IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS60(9) · pp.3814 · 2013
Citation
Web of Science 47
JCR IF
JCR 2012 5.165
ISSN
0278-0046

Authors

  1. Lee, JaeohFirst authorDepartment of Electronics Engineering, Pusan National University, South Korea
  2. Han, Seong IkCorresponding authorDepartment of Electronics Engineering, Pusan National University, South Korea
  3. Lee, JangMyungCorresponding authorDepartment of Electronics Engineering, Pusan National University, South Korea

Abstract

This paper proposes a decoupled control algorithm for a single-wheel (unicycle) balanced robot. A unicycle robot is controlled by two independent control laws: the method involving mobile inverted pendulum control for the pitch axis and the method involving reaction-wheel pendulum control for the roll axis. We assume that both roll dynamics and pitch dynamics are decoupled from each other. As a result, the roll and pitch dynamics are obtained independently and all interactions between them are considered disturbances. Each control law is implemented by an individual controller, i.e., fuzzy-sliding mode control for roll and linear quadratic regulator control for pitch. Fuzzy logic is utilized to compensate for the interactions between the pitch and roll dynamics in real time. The unicycle robot has two dc motors: one to drive the disk for roll and the other to drive the wheel for pitch. Since there is no force to change the yaw direction, the dynamics of the yaw direction is not changed in this paper. Algorithms for the decoupled dynamics were implemented in a real unicycle robot, which was made to follow a desired trajectory along a straight line. Angle data was obtained by fusion of the gyro sensor and accelerometer. The results of experiments conducted confirmed the effectiveness of our proposed control system.

Keywords

Decoupled dynamicsfuzzy logicfuzzy-sliding mode controllinear quadratic regulator (LQR) controlunicycle robotWHEEL PENDULUM

Identifiers

WoS UID
000318702200025