Strong tracking filter based SINS initial alignment for large misalignment angles

Initial alignment is important for strapdown inertial navigation system. Coarse alignment is usually implemented before fine alignment to make the misalignment angles sufficiently small to meet the linear error model. Without coarse alignment, the misalignment angles will be large enough to need non-linear model and non-linear filter, which are complicated and inaccurate. For this reason, a new method is presented which don't need coarse alignment or non-linear model. This is achieved by choosing appropriate initial error covariance matrix and using feedback compensation structure and strong tracking filter. Simulation shows the new method is effective and robust for large misalignment angles. The heading error can ultimately converge to 0.05° after 300 s under arbitrary initial attitude, while the traditional fine alignment has almost the same accuracy and speed but needs additional time for coarse alignment, and the UKF only converges to 0.5° after 600 s which is much slower. © 2015, Editorial Department of Journal of Chinese Inertial Technology. All right reserved.