Gaussian-Binary classification for resident space object maneuver detection

The capability accurately and timely detect a resident space object (RSO) maneuver is critical task for monitoring activities. This paper presents data-driven Gaussian Binary RSO Maneuver Detection (GaBRSOMD) method to whether or not there between two tracks of the same on different orbital paths. Using an in-house simulated catalog environment, Classification (GBC) used three types maneuvers: due small impulsive velocity change; low thrust, either thrust which unknown priori. Numerical results demonstrate that proposed GBC model can achieve high accuracy in detecting all cases maneuvers. further demonstrates approach robust noisy data has advantages over AutoEncoder classical Principal Component Analysis (PCA) method. Hence, great potential making quick detection decision with both precision.