Improved attitude estimation accuracy by data fusion of a MEMS MARG sensor and a low-cost GNSS receiver

Abstract

To obtain precise and continuous attitude information, this paper proposed a data fusion scheme that exploits the complementary advantages of a microelectromechanical system (MEMS) magnetic, angular rate, and gravity (MARG) sensor and a low-cost global navigation satellite system (GNSS) receiver. A quaternion-based error state Kalman filter is designed to integrate the different data sources, where the attitude error and gyro bias variation are taken as the error states. The gyro-measured angular rates are continuously integrated to propagate the attitude quaternion, while the accelerometer and magnetometer readings or the GNSS pitch and heading angles estimated by the precise carrier phase observations are utilized to perform the measurement update. The system kinematic model remains constant when the aiding source changes, only the measurement model alternates accordingly. A vehicular dynamical test was conducted to evaluate the proposed algorithm, the test results show that the attitude estimation accuracy of the integrated algorithm is apparently improved.