In-motion coarse alignment method based on reconstructed observation vectors

Abstract

In this paper, an in-motion coarse alignment method is proposed based on the reconstructed observation vectors. Since the complicated noises are contained in the outputs of the inertial sensors, the components of measurement observation vectors, which are constructed by the sensors' outputs, are analyzed in detail. To suppress the high-frequency noises, an effective digital filter based on the Infinite Impulse Response technology is employed. On the basis of the parameter models of the observation vectors, a new form Kalman filter, which is also an adaptive filter, is designed for the recognition of the parameter matrix. Furthermore, a robust filter technology, which is based on the Huber's M-estimation, is employed to suppress the gross outliers, which are caused by the movement of the carrier. Simulation test and field trial are designed to verify the proposed method. All the alignment results demonstrate that the performance of the proposed method is superior to the conventional optimization-based alignment and the digital filter alignment, which are the current popular methods.