Automated detection and characterization of earth-orbiting satellite maneuvers

Abstract

The process validates measurement data that are consistent with the current orbit or allowable maneuvers for Earth-orbiting satellites. The validated data may be processed to detect both large and small maneuvers; large maneuvers that may have been rejected by standard data validation process and small maneuvers that may have been missed. Large maneuvers may present as 'trends' in the residuals above the noise threshold from multiple sensors. Small maneuvers may present as a deviation from the expected orbit. The OD engine is suitably configured (as per normal operation) to enforce data reduction on any outliers (e.g. >3 sigma). The maneuver estimation routines employ an objective function for selection of maneuver estimates that combines a residual fit to the current orbit and maneuver, orbit consistency and a penalty for validated but unused data (data that is rejected by the OD engine). The combination of enforcing data reduction in the OD engine of validated data while penalizing that reduction biases the estimation routines to select maneuver estimates that fit all or substantially all of the validated data. The maneuver estimation routines also suitably use a center time & duration window to select the maneuver estimate and iterate the process as new measurement data is received to refine the maneuver estimate.