A Fast Two-Position Alignment Method Based on Pseudo-Attitude Measurement for Low-Cost SINS

Large azimuth error has been a major conflict in static alignment, and the accuracy of initial alignment is closely related to the performance of inertial measurement unit(IMU). In particular, the azimuth error suffers from the equivalent east gyro drift. To improve the precision of attitude estimation as well as IMU calibration, a fast two-position alignment scheme is proposed which mainly targets low-cost IMUs. Based on the attitude invariant constraints under static, a pseudo attitude observation with fixed error is introduced at the first position. Meanwhile, the state equations are extended and a segmented measurement model based on velocity and pseudo-attitude is constructed. In addition, the optimality of the scheme is theoretically supported by combining the analytical approach and improved observability analysis. Only one rotation maneuver is required to decouple all states and maximize the observability degrees as well as estimation accuracy. By comparison simulations with conventional multiposition alignment schemes and real experiments, the advantages of the proposed method are demonstrated in terms of high accuracy, good stability, fast convergence, and easy operation.