Research Of The Transfer Alignment Error Compensation And Precision Estimation

Before the SINS (slave inertial navigation system) is entirely aligned, the difference between MINS (master inertial navigation system) and SINS performance parameter output can reflect the size of misalignment in some extent. Make use of this theory set up corresponding error model and using some estimate method, then the misalignment can gain.This paper's researches mainly based on actual using of transfer alignment; firstly, the common matching methods was introduced particularly, and then have studied the compensation methods of the Lever Arm effect, flexibility distortion and transfer delay deeply. The last party of this paper is mainly focus on the research of transfer alignment precision estimation.Firstly, use velocity matching as the research object, and have researched the theory of Lever Arm effect and its effect to transfer alignment in this paper. Then this paper has designed two methods to compensate the Lever Arm effect, one is using digital filter, another one is compensating by computing directly. About the compensation of flexibility distortion, this paper using second order markov process as the model of flexibility distortion, and make the flexibility distortion as party of the system's states, the result showed that this compensation method can gain a good purpose. To transfer delay, also use transfer delay as party of the system states, and give out the new observation equation under the transfer delay. Through simulation the result shows that this method can eliminate the effect of the transfer delay in transfer alignment.Secondly, another important part of this paper is precision estimation of transfer alignment, in this paper precision estimation error model of strapdown inertial navigation system was gave out, use kalman fixed point smoothing and fixed interval smoothing method to smoothing estimate the alignment precision. It is useful proved by simulation. In the last part of this paper, the half practicality simulations of velocity matching and velocity plus attitude matching were carried out, and the result is satisfactorily.