Orbit Determination Performance Improvements for High Area-to-mass Ratio Space Object Tracking Using an Adaptive Gaussian Mixtures Estimation Algorithm
- Paper number
IAC-09.A6.2.1
- Author
Dr. Moriba Jah, Air Force Research Laboratory (AFRL), United States
- Coauthor
Dr. Thomas Kelecy, Boeing Integrated Defense Systems, United States
- Year
2009
- Abstract
Inactive high area-to-mass ratio (A/m) resident space objects (RSOs) in the geosynchronous orbit (GEO) regime pose a hazard to active GEO RSOs. The combination of solar radiation pressure (SRP) and solar and lunar gravitational perturbations cause perturbations in the orbit parameters of the inactive RSOs. The high A/m nature of these objects results in greater sensitivity to SRP resulting in the perturbation of mean motion, inclination and eccentricity. The subsequent drift with respect to Earth-based tracking sites, combined with time varying orientation with respect to the sun, results in many of these RSOs being “lost” after initial acquisition as they transition through periods of days to weeks out of view of observing sites. The unknown material make-up, approximations in the shadow function used for eclipse modeling, errors in size estimates and variations in the orientation with respect to the sun result in unpredictable (random) errors. Analysis and simulation results show the distributions of propagation errors resulting from each of the observation and modeling uncertainties to be non-Gaussian. Figure 1 shows the mean total position error distribution for a A/m = 10 m2/kg GEO object propagated over 1-day, where the errors include un-modeled rotation, un-modeled atmospheric refraction and absorption of the solar rays during a 1-hour eclipse, a 1% error in the estimated A/m and 1% errors in the specular and diffuse reflection coefficients. The distributions of the position errors from the individual error contributions are seen to be quite different as, for example, in Figure 2 where the total position error distribution is shown for A/m estimation errors only. Results of recent work that has been done (Terejanu, et al., 2008 ) show promise when an adaptive Gaussian mixture approach is applied to the determination of attitude in the presence of attitude observation and modeling errors. The appropriate SRP models are incorporated into an adaptive Gaussian mixtures implementation of an orbit estimation algorithm, and performance improvements to the estimated state are assessed. Sensitivity to materials, shadow model and orientation modeling errors in the SRP acceleration computation are included. This is a way to improve on the orbit estimation by applying a technique which does not rely upon the Gaussian assumptions that are often applied to non-linear estimation techniques. The by-product of more realistic covariance information should improve the accuracy of conjunction assessments. The goal of this study is to define the orbit processing and prediction requirements for the tracking of high A/m RSOs using an adaptive Gaussian mixture approach. Figure 1. Total Position Error Distribution for a High A/m GEO Object: Combined Rotation, A/m, Specular and Diffuse Reflection Modeling Errors Figure 2. Total Position Error Distribution for a High A/m GEO Object: A/m Estimation Errors Only
- Abstract document
- Manuscript document
(absent)