基于奇异谱分析的空间非合作目标无源机动检测方法

To solve the maneuver detection problem in the initial orbit determination stage of the space situation awareness mission for non-cooperative targets, a passive maneuver detection method based on singular spectrum analysis is proposed. Firstly, a nonlinear relative dynamics and a line-of-sight measurement model are established in the local vertical local horizontal frame. Subsequently, the features of the line-of-sight (LOS) angle sequences are extracted by singular spectrum analysis, the calculation method of the differential evaluation index and threshold are designed, and the AND/OR gates are used for comprehensive judgment. Finally, the maneuver detection model is verified, tested and analyzed in GEO type orbit and different working conditions. With a camera precision of no less than 3×10^-4 rad, small pulse maneuvers within the range of 1 m/s for non-cooperative targets with an altitude difference of 1000 km could be effectively detected, the success rate of maneuver detection above 1 m/s is more than 90%.