Robot Attitude Estimation Based on Kalman Filter and Application to Balance Control

To solve inertial sensors in the presence of random error for robot balance control, this paper proposes a simple and practical Kalman filter (KF) algorithm, which was implemented to information fusion for inclinometer and gyroscope, thus get optimal estimation for the robot attitude signal after sensors error compensation. The experimental results showed that the method based on KF to obtain the optimal estimation is effective and feasible. Through the simulation including balancing control experiment and disturb experiment, the two-wheeled robot by using LQR method can keep balance in fixed position well by the proposed method. Comparing to the method not using KF we can verify the feasibility and effectiveness of the KF attitude estimation method. The robot has a better ability of self-balancing control after using KF.