Deep Space Autonomous Orbit Determination Using CCD

In this paper an algorithm for orbit determination in deep space using only measurements of attitude and of line of sight of known celestial bodies is presented. Measurements are then processed using two different filters reconstructing part or the entire state vector: a batch least squares filter and a combination of batch and Kalman filter. The strategy and the effectiveness of the two filters are compared on two sample cases of interplanetary trajectory toward Mercury (inner part of the solar system) and toward Jupiter (outer part of the solar system). Introduction One of the driving issues for near future missions is autonomy: in fact autonomy reduces the need for support from ground, so reducing ground station development and maintenance, cost, man effort, reaction time to unexpected situations, and, avoiding dangerous delays in control loops, it is an essential concept when dealing with cheap and multiple satellites and very long missions, where long time is required for transfer. Historically navigation of deep-space satellites has involved the use of radio data types for determining trajectory of the spacecraft and then predicting its course in the future. These radio data types include Doppler data to determine the component of the line of sight velocity and ranging to determine the line of ✫ Research Fellow. ✞ Full Professor Graduated Student Copyright 2002 by the authors. Published by the American Institute of Aeronautics and Astronautics,Inc., with permission. Released to AAS/AIAA to publish in all forms. sight position. Doppler measurements are acquired on Earth in S and X band to determine the velocity of the spacecraft while position can be computed measuring the transit time of a radio signal from the ground station to the spacecraft. Therefore traditionally navigation is completely demanded to operations on Earth, with a consequent cost in terms of human resources and to the fundamental issues related to the delay in receiving and sending signal toward and from the spacecraft or related to possible occultation or failure that may jeopardise the entire mission. For some missions, during approach to target bodies, optical data taken with an onboard camera were used to measure the target-relative position of the spacecraft, providing additional information. This methodology of combining both radio and optical data has worked very well in the past (Voyager and Galileo missions). This can be regarded as an early attempt of onboard autonomy especially for Galileo. The same concept for image processing, experimented on Galileo, was implemented onboard the first mission designed to test autonomous navigation systems: Deep Space 1. In fact. for purposes of developing autonomously navigating spacecraft, the software needed to process radio navigation data is very complex. In addition a radio system onboard cannot be considered really autonomous because an uplink from the ground is needed. By its nature, optical data are not as precise as radio systems, but have the distinct advantage of being self-contained onboard the spacecraft. DS-1 system was enriched with a completely autonomous centerfinding procedure of the objects used to determine its position and, as a consequence, the position of the spacecraft using an ephemeris database. In this paper a strategy for orbit determination based on measurements of LOS (Line Of Sight) of known celestial bodies is presented. Two algorithms