An Improved Double-Layer Kalman Filter Attitude Algorithm For Motion Capture System

Zequan Zhang,Qi Jin,Wenguang Jin

DOI: https://doi.org/10.1109/CISP-BMEI51763.2020.9263618

2020-01-01

Abstract:The improvement of measurement accuracy for motion capture systems has been drawing the great concern of researchers. And attitude angle plays as a key factor in measurement accuracy, which is obtained from the multi-sensor fusion. This paper proposes an improved attitude algorithm, that is, a combination of the complementary filter (CF) and the double-layer Kalman filter (DLKF) algorithm. We get the attitude angle after a complementary filter and take it as the observation variable, and then uses a double-layer Kalman filter for data fusion. Meanwhile, to avoid magnetic interference, the least square method(LSM) is taking to fit the magnetometer data, as well as avoid magnetic interference. To verify the feasibility and effectiveness of the algorithm, a self-designed MPU9250 sensor-based inertial measurement unit (IMU) was used for testing. The results show that the algorithm can realize accurate attitude angle calculation with better suppression of noise and drift effects, and have better accuracy in calculating the yaw than the Mahony algorithm.