Визначення змін характеру руху об'єктів ракетно-космічної техніки за авторегресійними моделями в умовах нерівновіддалених спостережень

Abstract

The problem of increasing prediction accuracy for the motion of Earth-orbiting objects (EOOs) and detecting changes therein is topical for the tasks of spacecraft life prediction, space debris cataloguing, and navigation. Therefore, the problem of detecting changes in dynamic systems characterized by non-equidistant observations is topical. The purpose of this work is the development of autoregressive models with observations non-equidistant in time to detect changes in EOO motion. The methods employed are multivariate statistical analysis, time series prediction, and complex-system simulation under structural uncertainty. Data generated by NORAD (USA) were used as initial observations to describe EOO motion. They are actual, constantly updated, and freely available via the Internet. These data are presented in the Two-Line Element (TLE) format, which is a data format encoding a list of orbital elements of an EOO for a given point in time. This paper presents a method for constructing autoregressive models to describe the dynamics of EOOs represented by time series of TLE elements with values non-equidistant in time. On its basis, autoregressive models of the Sich-2 spacecraft’s dynamics were constructed. The standard errors of the models were analysed on examination samples, and significant deviations of the standard errors for the basic variables (apogee, perigee, eccentricity, longitude of ascending node, perigee argument, and average anomaly) were found, thus demonstrating changes in the Sich-2 motion from its basic regime. The novelty of this work lies in that the problem of detecting changes in EOO motion characteristics based on the proposed type of autoregressive models has not been considered before. Its practical value lies in that the simulation of the Sich-2 motion using time series of TLE elements allows one to detect changes in motion regimes; the method may be used in detecting in-service changes in EOO properties.