Application of Fuzzy Information Intelligent Computing Theory in Space Target Recognition

Abstract

In this paper, the fuzzy relative entropy of fuzzy information theory is introduced for the spatial target orbit information and its related characteristics are discussed. For the tens of thousands of spatial target orbital feature information in the space target orbit information database, the J2000 inertial mean orbit elements are used to construct the respective fuzzy feature vectors. The fuzzy relative entropy discriminant algorithm is used to model and analyze the big data information of the target and the orbit information base. After 10 sets of near-ground circular orbits and 6 sets of medium-high orbits to be identified and fuzzy relative entropy recognition calculation of the orbit information database, the proportion of the target to be identified correctly identified as a unique spatial target can reach up to 63%. Due to the influence of the orbital error and the threshold error, the initial identification of the orbit is often as many as several dozen. Compared with the traditional empirical threshold interval recognition technology that relies on the orbit feature information, the method can greatly improve the success rate of the recognition calculation. Since the spatial information of the target orbit feature itself is fuzzy, there is always a deviation from the actual rail orbit calculation. Therefore, the fuzzy relative entropy method is reasonable. The method applies the information discrimination characteristics of fuzzy information theory to measure the spatial target orbit. Compared with the traditional orbit feature information recognition method, it has the characteristics of high computational efficiency, large information discrimination, strong anti-interference performance and high recognition success rate. It has a good application prospect in the field of space target (fragment) detection and identification.