Modeling non-deterministic behavior in complex systems
The Dutch National Aerospace Laboratory proved that advanced software tools written in the Unified Modeling Language can successfully model the complexity that plagues reconnaissance aircraft. Medium Altitude Reconnaissance (MAR) aircraft are used for military as well as civilian observation activities. MAR image collection is highly challenging technically in comparison to geostationary satellites due to the fact that the aircraft is moving in relation to its target. Yet MAR images provide much higher resolution. Scientists with the National Aerospace Laboratory (NLR) in the Netherlands took part in a MAR case study in the framework of an Information Society Technologies project. Their goal was to learn if new software tools developed with the Unified Modeling Language (UML) by academic partners in the OMEGA project could be applied to improve MAR performance. At the heart of the experiment is the MAR's avionics data bus, which is responsible for monitoring critical data from a number of different systems. The situation is complicated by the fact that data is supplied at both regular and irregular (asynchronous) intervals. The ability of the UML Verification Environment (UVE) tool to model the inherent non-determinism allowed the NLR to determine if data loss negatively impacted safety and other factors. Correspondingly, the IF/IFx tool facilitated the creation of an accurate model of the timed system components, which has been further validated by user input. Hence, the case study demonstrated the usefulness of both the timed and untimed system verification tools developed during OMEGA. Future research planned by the NLR and its OMEGA partners will focus on the effect of the new tools on the overall software development lifecycle in the MAR and other case study domains.
