Taking the ECSS Autonomy Concepts One Step Further
Fabricio de Novaes Kucinskis, Mauricio Goncalves Vieira Ferreira · SpaceOps 2010 Conference · 2010
This paper presents the first efforts made at the Brazilian National Institute for Space Research (INPE) to increase the autonomy of its satellites through the adoption of model-based reasoning on the flight software. A prototype for an on-board, goal-oriented replanning system was developed. With the lessons learned from this prototype, we are implementing an on-board service that provides states inference, which can be used as an autonomy kernel by autonomous applications. But instead of creating one more brand new, one-of-a-kind system, we are trying to make this service fit into the European Committee for Space Standardization (ECSS) standards, which have been adopted by INPE. ECSS has not only defined an autonomy concept for space missions, but has also divided it in categories and levels. We identified a gap between two of the last autonomy levels, and propose a new intermediate level to fill it. Our Internal State Inference Service (ISIS) implements this new level. I. Introduction It is a general agreement among the space community that spacecrafts have to be increasingly more autonomous. A survey 1 on NASA’s command execution systems highlights that “as NASA’s missions become more complex, autonomy becomes increasingly important to the success of those missions.” A paper about ESA’s PROBA satellite 2 states that “it is generally expected that, in the near future, mission operations activities will evolve from ground-based control towards on-board monitoring and control”, and that “more complex functions may gradually be migrated to the on-board computer. On-ground operations spacecraft activities by humans will essentially be limited to the initialization phase (establishing the routine operations) and to emergency support”. But the transfer of new functions to the on-board computer is a hard task. It involves the capturing of a large amount of knowledge from human experts, the filtering of the resulting knowledge base to keep just what is needed for the mission to be accomplished, and the representation of that knowledge in a concise and unambiguous form. Moreover, the flight software (FSW) has to operate over the knowledge base with scarce on-board computational resources, and make decisions in such a way that the mission is not endangered, nor its performance compromised. That's why, except for natively closed-loop tasks like attitude control, only few missions have flown some kind of autonomous FSW – almost all of them for limited periods of time, to perform one-shot technology validation experiments. However, space agencies such as NASA, ESA and CNES keep carrying research and developing prototypes for autonomous systems. The Brazilian National Institute for Space Research (INPE, in the Portuguese acronym) is also trying to address these problems. We have developed a prototype for a model-based autonomous replanning FSW, and this first experience allowed us to try a wider approach, an ‘autonomy kernel’ in the form of an on-board service. This service provides states inference, resources profiling, constraint propagation and simple temporal networks features for onboard autonomous applications that perform tasks such as diagnosis/prognosis and goal-oriented mission replanning. But to make this service ready to flight in our future satellites, we decided to avoid the development of one more brand new, one-of-a-kind system and are trying to make it fit into the European Committee for Space Standardization (ECSS) engineering standards, which have been adopted by INPE. This paper presents our work and how it relates to the ECSS standards. Chapter II has a compilation of the ECSS autonomy concepts and the related engineering standards. Chapter III shows a survey on the levels of autonomy reached by the current INPE missions. The efforts that have been made at INPE to increase the FSW autonomy for its future satellites are reported on Chapter IV. Chapter V brings our proposition for an extension of the ECSS autonomy levels. An Internal State Inference Service that meets this proposition is described in Chapter