ESA GNC Conference Papers Repository

Title:
Real Time Performances of a Vision Based GNC for Automated Rendezvous in Elliptical Orbit
Authors:
Bastante, J.C. ; Vasconcelos, J ; Monteiro, J ; Dinis, J ; Hagenfeldt, M ; Salehi, S.
Presented at:
Karlovy Vary 2011
DOI:
Full paper:
Abstract:

Next future exploration missions call for the maturation of the associated key technologies enabling the implementation of several of the mission phases in a reliable, robust and optimized way. The maturation of such technologies, measured through the TRL scale, requires the conception of not only simulators, but also a development framework in which the validation of the GNC algorithms can evolve in sequential steps of increasing representativeness with respect to a real, operational implementation of those algorithms. One of these technological fields is automated Rendezvous, in either circular or elliptical orbit, which is indentified as enabling for sample return like missions. During several years, DEIMOS has been designing a complete framework for the conception, first, and the sequential validation, after, of the GNC algorithms needed to run a complete end-to-end RV profile for capturing a sample in a Martian environment and in an elliptical orbit. The only considered sensor for relative motion is a camera taking images of the target. This framework is composed of three main simulators: a RVD-FES (Functional Engineering Simulator), fully coded in MATLAB/Simulink and used to test the functional performances of the GNC algorithms; a RTTB (Real Time Test Bench) with only Processor in the Loop (PIL), which after autocoding, tests the GNC algorithms running in a representative processor board (LEON3, in this case); and finally, a RTTB with also Hardware in the Loop (HIL), in which GNC and Image Processing (IP) algorithms are integrated within the LEON processor. IP is fed by images taken by a camera to a physical target in laboratory. The laboratory mock-up is designed to close the loop in real time in what regards the measurements (images) of the relative motion, while GNC and IP algorithms run in real time integrated in the LEON3 processor.