Experiment of a Quaternion-Based Extended Kalman Filter with Noise Reduction Technique Using MEMS Sensors for Attitude State Estimation

Position and orientation acquirement is one hot topic in sensor data fusion research and application area. Micro-Electro-mechanical Systems (MEMS) sensor is a kind of low-cost and feasible tool that could measure some movement information, including acceleration and angular rate and so on, while burst noise and outlier signal influence the sensor data for accurate information acquirement and related application analysis. Some advanced digital signal processing techniques could realise the improvement of signal quality. The algorithm of median filter, combined with mean filter, was implemented to output better data for orientation estimation with fewer burst noise and outlier signals. Considering our designed and developed sensor system, including a tri-axis accelerometer, a dual-axis pitch and roll gyroscope and a yaw gyroscope used in this research, the construction of sensor model and a kind of quaternion-based extended Kalman filter (EKF) method and involved formulas for attitude state estimation are described in this paper. The experiment of this quaternion-based EKF with noise reduction processing was implemented using MEMS sensors. Regarding the EKF result of orientation estimation, this paper analyses the reason why the yaw drift generated. The EKF result of orientation estimation proves the feasibility of 2 degrees of freedom for attitude state estimation application.