Fuzzy based tuning of a sensor fusion based low cost attitude estimator

A fuzzy based tuning mechanism of the Kalman filter (KF), for the purpose of attitude estimation using low cost Micro-Electro-Mechanical System (MEMS) sensors is proposed. The conceived system is fitted with three axis gyroscope, accelerometer and magnetometer as the sensor suite. The filter's dynamics is based on a quaternion based attitude propagation model using rate gyros; with the other two sensors providing a complementary attitude estimate. The measurement and process covariance of the filter are tuned using a Takagi-Sugeno based fuzzy logic model based on the perceived system dynamics and filter divergence indicators. The performance of the developed system shows better attitude estimation and convergence [9]properties than a conventional KF based attitude estimation system.