A Method for Rapid Transfer Alignment Based On UKF

Using quaternions is an effective approach for solving stiff numerical problem in the transfer initial alignment using Eulerian angle attitude algorithm when the missile is in vertical state. However, due to the strong nonlinearity, the usual Kalman filter can not be applied. Therefore, we use UKF to implement the estimation in nonlinear model. Based on UT transformation and adopting Kalman filter´s frame, UKF makes sampling strategy approach nonlinear distribution and the specific sampling form is determinate sampling. The number of sampled particles (usually named Sigma particles) of UKF is very small, which is determined by the selected sampling strategy. The most typical strategy chooses 2n + 1 (n is the extended state dimension) symmetrical Sigma particles. The amount of calculation of UKF is approximately equal to that of EKF, but, its performance is better than EKF´s performance, and the determinate sampling makes UKF avoid PF´s particle degeneration. This method based on UKF can implement initial alignment when the missile is in any attitude state. This paper presents the nonlinear model of initial alignment and uses UKF to implement the estimation. The simulation shows, this method can obtain high precision in short time