effective orbit determination following multiple spacecraft separation

Paper number

IAC-08.C1.4.4

Author

Mr. Fabrizio Reali, Aerospace Engineering School, University of Rome "La Sapienza", Italy

Coauthor

Dr. Giovanni B. Palmerini, Universita' di Roma 'La Sapienza', Italy

Year

2008

Abstract

Orbit injection of clusters of satellites on board of the same launcher is nowadays a current practice to reduce costs and increases flexibility in launch manifesting. If, as a frequent case, the final orbit is similar for all the platforms, the separation of the different spacecraft is accomplished in a limited time, the following identification of the spacecraft, and the relevant orbit determination, is not a trivial issue, and ambiguity problems arise.

The paper proposes to introduce, in such a specific frame, recent findings by the estimation theory. Following the widespread success of the classical Kalman algorithm in both the linear and extended versions, several improvements, like the Unscented Kalman Filter or the Particle Filter, have  attempted. All of these modifications could (have) be seen as prediction/update of a probability density function (pdf) associated with the phenomena under scrutiny. Following this approach, with the pdf, more than the kinematics state, at the centre of the scene, an additional, purely statistical level can be superimposed to the investigated dynamics, defining the identification of a specific process among several, possibly similar, others. Such a level will be represented by another pdf, variable in time and therefore being updated during the process, assessing the attribution of the measurements to the different targets. Coming down to the applications, it means that the dynamics of a single target (spacecraft) can be estimated based on the measurements gained form the global set. The behaviour of each single spacecraft will be modelled as the combination of a bulk motion, common to all the target, and of a residual specific to each platform. Each single measurement will be assigned to a target according to the relevant probability function, which will reflect the possibility that the corresponding residual motion, once measurement and process noises have been added, could result in the observed data.

The results of this non traditional approach, known as dense filtering, look quite interesting in terms of handling of the ambiguities in observation process, and lead to globally safer estimate. The paper aims to show the advantages of such a technique for orbit determination immediately following the separation, where ambiguity is a big concern. For the case of a satellite formation launch, the examples will be carried on until the acquisition of the final configuration, in order to show the properties of dense tracking all along the acquisition and operational phase of a crowdy mission concept.

Abstract document

IAC-08.C1.4.4.pdf

Manuscript document

IAC-08.C1.4.3.pdf (🔒 authorized access only).

To get the manuscript, please contact IAF Secretariat.