Quaternion-Based Kalman Filter for Micro-machined Strapdown Attitude Heading Reference System

AbstractA Kalman filter used in strapdown AHRS (Attitude Heading Reference System) based on micro-machined inertial sensors is introduced. The composition and principle of the system are described. The attitude algorithm and error model of the system are derived based on the quaternion formulation. The real-time quaternion-based Kalman filter is designed. Simulation results show that accuracy of the system is better than 0.04 degree without disturbance of lateral acceleration and reduced to 0.44 degree with lateral acceleration of 0.1 g/1Hz. Primary tests show that the static accuracy of the system is + / − 0.2 degree for pitch and roll, and + / − 0.3 degree for azimuth