Attitude estimation algorithm based on triaxial angular accelerometer

Due to the low precision of the inertial component in MEMS inertial navigation system, the attitude solving error is large. In order to reduce the attitude solving error, a method for estimating attitude using angular accelerometer and gyroscope is proposed. The proposed method is based on the extended state vector. The system consists of an angular accelerometer and a low-cost MEMS gyroscope. Angular acceleration is added to the state vector, and the attitude estimation model is built. After filtering and initial estimation using the angular velocity from the integrated data of angular accelerometer, Kalman filter is utilized to achieve the data fusion with MEMS gyroscope data. Experimental results on a turntable demonstrate that the proposed algorithm can reduce the drift error, and the average absolute error is reduced by 43.48% ~ 55.52% compared with the attitude solving scheme using MEMS gyroscope only, which improves the attitude solving accuracy of low-cost inertial navigation system.