Model-based Verification and Validation for Procedure Authoring
Guillaume P. Brat, Dimitra Gannakopoulou, Michel Izygon, Emmy Alex, Lui Wang, Jeremy E. Frank, Arthur Molin · 2009
The ”Apollo 13” movie was a great account of how dangerous human space flight is. It clearly showed that there is a fine line between mission success and catastrophic failures. Besides being a great thriller (we all wanted the astronauts to make it back safely, even though things look really bad for a while), it also offered a great look at how things are ran at NASA. The movie clearly showed that all activities are planned in the most minute details and described in procedures. It was true then, and, it is still true now for the Space Shuttle and the International Space Station (ISS). Procedures are plans for crew (i.e. astronauts) and flight controllers (which provides guidance from the ground). There are literally thousands of them for the ISS and the Shuttle; they will also be used on the new vehicle, called Orion, being developed by NASA. They may have to be adapted for different situations. The movie actually showed a dramatic example in which a power-up procedure had to be adapted so that the power load stayed under a certain amperage. That procedure was tried and (sort of) validated in a flight simulator before being given to the crew. The interactions between sub-systems were so subtle that the procedure could not get worked out on-board by the crew. It needed to be carefully adapted and validated on the ground. While technology might have evolved since the Apollo area, system complexity has certainly not decreased. Therefore, procedure authoring, validation and verification are still highly critical activities. This paper describes our current effort in improving procedure authoring, and more specifically, what can be done to speed-up and improve their verification and validation (V&V). We start by justifying the need for better procedure V&V. In the next section, we describe our example, the power system for the ISS. We then present the A4O (Autonomy for Operations) project, under which these technologies have been developed. The following section describes our analysis and the challenges we encountered during the analysis. Finally, we conclude with some lessons learned and present our future work.