A Comprehensive Survey on Linear Quadratic Regulator Control for Unicycle Robots: Experimental Insights

Abstract

This paper presents an experimental study on the application of Linear Quadratic Regulator (LQR) control for stabilizing a unicycle robot, or one-wheel mobile robot, a prominent example in the unicycle-type mobile robot group. Although inherently unstable, it offers several advantages over multi-wheeled, statically stable robots, such as requiring less space due to its single point of ground contact. The unicycle robot features two axes representing the two models used for analysis: the roll axis, modeled as an inverted pendulum controlled by the wheel, and the pitch axis, modeled as a reaction wheel inverted pendulum controlled by a reaction disk. LQR technique is a great method for developing a controller for nonlinear systems. Our approach involves designing, implementing, and evaluating LQR controllers with varied weighting matrices to assess their impact on system performance. Case studies are conducted by adjusting the weighting matrices of the LQR controllers and comparing each configuration with the initial LQR. Experimental results provide insights into the effectiveness of LQR for achieving robust control of unicycle robots.