Application of Expert Systems to Enhance On-Board Training in Long Space Manned Missions 941322
This paper focuses how a tool based on Expert System technology could best meet in orbit crew training requirements specifically in the conditions of long in-orbit stay.
The following important aspects are in particular addressed:
Man Machine Interface (MMI);
apportionment of capabilities between orbit and ground facilities;
possible architectural solutions for its use in a manned space module, according to the specific constraints of the orbital environment;
evaluation of available technologies for their possible integration in the set of the tools candidate to support crew member during their orbit stay time.
Emphasis is given to possibilities offered by state-of-the-art methodologies, w.r.t. knowledge acquisition, representation and maintenance, in order to provide astronaut with quick and context-sensitive information about training tasks to be performed, both planned and contingent.
It is shown how such a knowledge-based approach allows the utilisation of trainers' knowledge on astronauts (gained during the on-ground training sessions) regarding, for instance, the learning curves of each crew member; their single specific skills and qualifications etc., and how it is possible to tune an ad-hoc rapid on-board training course able to optimise the required intervention time in case of failure, contingency and new payload operation.
A hybrid approach is described, i. e. based on both:
object oriented programming paradigm, to represent models of payloads, subsystems, task hierarchic structures, etc…;
production rules, to represent procedural knowledge, rules of thumb, etc…
Finally, identified evolution of the technologies in this peculiar application field are outlined so as to recommend diverse options for future space environments, where astronauts have to live and work at the best possible conditions.