Ontology for Objective Flight Simulator Fidelity Evaluation

Abstract

The term simulator fidelity has become enor- mously important in the scope of simulation research, when assessing training efficiency and the transfer of training to real flight. It is defined as the degree to which a flight simulator matches the characteristics of the real aircraft. Objective simulator fidelity provides an engineer- ing standard, by attacking the fidelity problem with com- parison of simulator and the actual flight over some quan- titative measures. Research flight simulators encompass some differences from commercial flight simulators. They require high flexibility and versatility concerning the cock- pit layout and visual and motion systems, as well as flight simulation models. It should be easy to modify the flight simulation model or other software and hardware compo- nents of the simulator. To support this, there is a need for a flexible automated test methodology, in order to deter- mine the fidelity of the most relevant simulator subsys- tems, since they are often modified during the life cycle of the simulator. This methodology not only shall support automated execution but also enable automated genera- tion of the test cases which are subject to change as well as simulator components. The Institute of Flight Systems (FT) at the German Aerospace Center (DLR) has a recon- figurable flight simulator, the Air Vehicle Simulator (AVES), for research of rotorcraft and fixed-wing aircraft. The study reported in this paper adopts a Model Based Testing approach to tackle the high flexibility requirement of AVES. The outcome of the paper is a metamodel for model-based objective flight simulator evaluation. Meta- modeling has been carried out in two levels. An Experi- mental Frame Ontology (EFO) has been developed adopt- ing experimental frames from Discrete Event System Speci- fication (DEVS), and as an upper ontology to specify a formal structure for a simulation test. Then in Objective Fidelity Evaluation Ontology (OFEO) that builds upon EFO, domain specific meta-test definitions are captured.