Pedestrian Positioning Using A Quaternion-Based Indirect Kalman Filter

Error modeling and correction is one of the important part of pedestrian navigation.In this paper,we focus on the SINS errors,derives relation models of the major SINS errors.By recombining the relation formulas,we develop a novel Kalman filter which estimates six AHRS errors including quaternion error,angular rate error,linear velocity error,distance error,acceleration error and geomagnetic error.Then traditional static detection methods which based on specific force magnitude and slip variance for pedestrian navigation are discussed.Some failure possibilities in traditional methods are revealed after the equation between specific force magnitude/slip variance and external acceleration being established.A new adaptive method using angular rate is developed,which is based on the gyroscope and accelerometer error model.Using the novel kalman filter and the new static detection method,a fresh inertial navigation algorithm is developed.Static and Dynamic experiments have been done using VN-100 MIMU.