abstract- papersupto4pagesshouldbesubmittedusingthis（抽象——papersupto4pagesshouldbesubmittedusingthis）（5页）.docVIP

Abstract - In order to solve the flexure compensation problem in rapid transfer alignment, the error equations are simplified by noise compensation method firstly. Due to the time variant characteristics of flexure process in time domain, which leads to the fixed noise statistical characteristics cannot follow the variation of actual environment, the noise estimator with limited memory index is proposed. By limiting the memory length of obtained data, too old historical data is giving up and the accuracy of online noise estimator is improved. The final simulation verifies that the method proposed have higher accuracy and faster convergence speed than conventional methods. Keywords - Flexural Deformation, Limited Memory, Noise Compensation, Transfer Alignment vMI I. INTRODUCTION Transfer alignment is an important technology of moving base alignment. In the process of transfer alignment, the maneuver of carrier is required and the running environment is getting complicated. As a result, the flexural deformation is one of the most important error sources[1-4]. There are mainly two methods for the compensation of flexural deformation, the model construction method and noise compensation method. Because of the high complexity of physical modeling, most researches are turned to the experimental modeling, which uses the markov process to describe the random flexural deformation. The flexure motion is modeled as the third-order markov model driven by the white noise through analysis of experimental data in literature[1]. On the base of above job, the markov process is further divided into high frequency mode and low frequency mode by literature[2]. In literature[4], the elastic deformation of aircraft wing is modeled as two-order markov process. The model construction method is realized by computing the Markov parameters based on the similarity principle of power spectral density[5-8]. However, modeling of Markov process will lead to rapid increase of state dimensio