UDC 004.032.2
ON-BOARD DATA CLASSIFIER MODEL OBTAINED FROM ROCKET AND SPACE EQUIPMENT VEHICLE ON THE BASIS OF IMPLICIT FEATURES OF TRAJECTORAL DEVIATIONS
S. V. Spitsyn, post-graduate student, RSREU, design engineer, JSC «SRC «Progress» – department of special design bureau «Spectrum», Ryazan, Russia;
orcid.org/0000-0001-6468-5043, e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
S. A. Tikhomirov, assistant professor, RSREU, senior specialist, JSC «SRC «Progress» – department of special design bureau «Spectrum», Ryazan, Russia;
orcid.org/0000-0002-3174-1536, e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
The paper considers on-board data classifier model obtained from rocket and space equipment vehicle on the basis of implicit features of trajectoral deviations. The proposed classification model is based on specific binary classification algorithm in terms of modern and perspective machine learning methodology algorithms such as decision tree and random forest. The aim of the work is to find a generalized algorithm for determining the occurrence of abnormal and emergency onboard situations in complex multiparametric objects (rocket and space equipment vehicle) based on trajectory data analysis. The trajectory data is obtained by polygon measuring instruments as part of general flow of onboard telemetry information during flight experiment. The paper presents a step-by-step process of constructing effective classifier model for normal and abnormal on-board states. The model is based on modern methods of data mining and machine learning, which are currently not widely used in technological processes of rocket and space industry.
Key words: information and measurement support of rocket launches, trajectory measurements, machine learning, system analysis, telemetry information.