Research on the position estimation accodometry-based for the two-wheeled self-balanced robot

Aiming at the anomaly events, such as the wheel-slippage, the travel over obstacle, the bump and so on, the odometry invalidates in position estimation. The accodometry method is presented to fuse the encoder data and the accelerometer data to estimate position. The effect of non-systematic errors on the robot position estimation is solved. The ill-effects of accelerometer inherent drift are eliminated. The position determination accuracy is increased. The experiment demonstrates the effectiveness of the accodometry approach. Experimental results show that the accodometry error is about four times smaller than the odometry-only error.