A design for two-wheeled self-balancing robot based on Kalman filter and LQR

Abstract

This paper introduces a method to control a two-wheeled self-balancing robot based on Kalman filter and LQR algorithm. It is incorporated with hardware design, signal processing and control algorithm design. In this system the angle signals from a gyroscope and an accelerometer, disposed by Kalman filter, are combined with LQR controller to accomplish the controlling of the robot. The purpose of simplifying Kalman filter is to make the signal processing using fewer system resources and reduce the burden of the system. A mathematical model was established according to the physical model. Based on the mathematical model, the LQR controller was designed for the attitude and position controlling of the robot. The experiment and simulation based on the prototype robot make it clear that the system could achieve a state of equilibrium and position tracking with a initial disturbance in a short time.