A new approach for attitude estimation of unicycle robot

Abstract

Attitude Estimation is critical for balance control of Unicycle Robot, Accelerometers and gyroscopes are always used for posture detecting, and unscented Kalman filter(UKF) is adopted as sensors information fusion algorithm. Because the accelerometer is sensitive to external vibration and non-gravitational acceleration, the results of attitude estimation are extremely inaccurate. In order to overcome this problem, through the deeply analysis to Kalman filter algorithm and coupled with a strong validation of experimental results, this paper proposes a new approach based on the UKF algorithm. The new algorithm compensates the external acceleration errors by adjusting the measurement noise covariance adaptively. The experimental results show that the extended unscented Kalman filter algorithm has a good effect for solving unicycle robot posture detecting problems and eventually gets the more accurate attitude angles.