Research/ IN-SPACE SERVICING
A Method for Optical Tracking of On-Orbit Servicing Operations in Geostationary Orbit Using Speckle Interferometry
Robert Scott
On-Orbit Servicing (OOS) missions in geostationary orbit with inter-satellite separations less than one kilometer pose a problem for ground-based electro-optical space surveillance sensors. The tight separations between space objects subtend angles comparable to the size of turbulence (seeing) cells of Earth’s atmosphere making classical ground-based space surveillance detection approaches problematic. Speckle interferometry using a cross-spectrum approach was explored in this research to overcome atmospheric turbulence and enable ground-based measurement of the relative positions of OOS satellites without use of adaptive optics. Field testing of this technique using a medium aperture telescope on co-located geostationary satellites, acting as OOS proxies, found that cross-spectrum speckle interferometry measurements can obtain in-track and cross-track relative position precisions better than 100 meters when the satellites’ apparent angular separations were less than 5 arcseconds. The cross-spectrum method was found to be susceptible to high fringe rotation rates inducing incorrect separation measurements during times when the co-located satellites were at their closest point of approach. While this effect produced spurious fringes during co-located satellite observation tests, this effect is not expected to be encountered during observation of a true OOS mission. Brightness differences did not pose a significant observational limitation as brightnesses of Mprimary = 10.2, Δm = 0.3 and Mprimary = 9.1, Δm = 1.5 were speckled successfully and relative position estimates obtained. Additional testing with small aperture telescope achieved similar brightness and positional measurement performance. Due to the plane-of-sky differential angle observations of the objects a partial observability condition is encountered where the radial component of the servicer’s motion requires more than three hours of observations to converge on a relative orbit solution. The near collinear alignment of the detector with the in-track and cross-track directions causes this partial observability condition. Fortunately, due to the dynamic coupling between the in-track and radial directions of relative orbital motion, the radial component can be estimated and the time to converge is largely dependent on the measurement precision.