Development of a coaxial self-balancing robot based on sliding mode control

A low cost coaxial self-balancing robot is proposed in this paper, the two wheels of which are placed coaxially for turning with zero-radius. Low cost MEMS accelerometer and gyro are selected to measure the posture information of the robot with a novel data fusion method. This data fusion method can overcome the shortcomings of accelerometer and gyro so that the precise posture information is obtained even with oscillation and impact. Based on the robot's dynamics model established by Lagrange's function method, two robust sliding mode controllers are designed for controlling the motions of the robot. Not only numerical simulation experiments using MATLAB Simulink and ADAMS but also physical experiments are conducted to confirm the effectiveness of the two controllers, and the results show that the robot performs well with precise measurement of the posture and sliding mode controllers.