A novel robust Kalman filter on AHRS in the magnetic distortion environment

This paper deals with the Attitude and Heading Reference System (AHRS) under the conditions of magnetic distortion. The conventional AHRS involves monitoring the difference between the measurements and the projections of gravity reference vector and geomagnetic reference vector in body coordinate (Right-Forward-Up). However, due to the magnetic distortion, the attitude estimated by AHRS will suffer from a fatal error. To solve this problem, we propose an indirect Kalman filter with two-step observations update, which makes the magnetic disturbance only influence the output of heading and not affect the estimated pitch and roll angles. In addition, the proposed system performs online detection of magnetic disturbances by exploiting variations in magnetic dip angle. According to the result of detection, we use the equivalent weight matrix to place more emphasis on the results of dead reckoning of the gyroscope measurements and less on the magnetic observation update. The experiment validation shows that the proposed system has an accurate and timely detection and has more robust performance than standard Kalman filter in the presence of magnetic disturbances.