Analysis And Research Of Pointing Error Of High Precision Multi-axis Stabilized Platform

High precision multi-axis stabilized platform keeps the line of sight(LOS) stable through isolating the carrier’s disturbance. It can hence capture, track and aim at the targets. Due to its superior performance, it has been widely used in civil, military and commercial applications. Based on the development of platforms in the domestic and overseas, this paper focused on its pointing accuracy. After the description of the platform’s structure and movement mechanism, it continued to analyze the error resources and corresponding propagation rules which affect the pointing accuracy. It also did a deep research on the error synthesis and distribution as well as test processes and compensation methods for relevant errors.This paper firstly gave a description of several commonly used gimbal structures and platform’s working mechanism. It obtained the platform’s location equations and the conditions in LOS stability by means of homogeneous transformation. As to various factors affecting its pointing accuracy, they were divided into two groups: static errors and dynamic errors. It continued with a detailed description of each error’s effect by means of quaternions. The paper final y established a specific error model, which had explicitly physical meanings.According to the theories of error synthesis and distribution, this paper has established each error’s probability density function, it analyzed relevant sensitivities by employing Monte Carlo method, which provided a guidance for error distribution. Further, the paper did some research on tolerance synthesis and distribution and gave a detailed description of relevant methods, followed by a specific example, which would guide the accuracy distribution in actual application.Moreover, the paper also analyzed and summarized commonly used calibration methods of the platform. One is based on a high precision turntable and an autocollimator. This paper conducted a deep research on the calibration process error by means of octonions, and final y obtained the process error model. It also solved the coupling problem in azimuth and elevation readouts.Error compensation as an effective method to improve the accuracy has been used in many aspects. This paper continued to establish nonparametric and semiparametric models to improve the platform’s pointing accuracy. Gaussian process regression based on Bayesian nonparametric estimation was presented in detail. An example was given out to compensate for the pointing accuracy by employing Gaussian process regression. Semiparametric model which combined advantages of both parametric and nonparametric models presented a better performance. This paper has established the semiparametric models based on penalized least square method and least square collocation method, respectively. And each model was detailed analyzed in terms of model estimate and prediction. During the comparisons of all the compensation models presented in the paper, semiparametric and nonparametric models presented a better performance.