Applying artificial intelligence theory to helicopter SAF simulation based on HLA/RTI. 2

Abstract

For pt.1. see IEEE Fall SIW, p. 333-40 (2000). This paper describes a modeling process and the federation results of the simulation of helicopter formation flight federation SAF (semi-automated forces) based on HLA/RTI. We also have realized an enemy entity that can move autonomously, thus, making this federation appear more realistic. The characteristics of this study are as follows. I) This formation flight model can conduct federation autonomously by applying artificial intelligence (AI) theory and a basic model of the human-decision process for helicopter flight. 2) We have developed a fundamental human-decision-process model for our helicopter formation flight based on insights gained,from responses to a questionnaire we used to get information from helicopter pilots about how they make decisions. This model can realize a federation that is more realistic than any other federation model. 3) In this model, if the command helicopter goes out of service, the other helicopter entities can carry on the mission autonomously. We demonstrated by a simulation that the result of a search mission does not change even if the command helicopter goes out of service. 4) An enemy can move in the battlefield and elude helicopter pursuit by using AI (which is like the constant satisfaction problem (CSP)) theory. 5) We have clarified the relation among the number of entities, the intricacy of a simulation scenario, and the network load on HLA/RTI. 6) We developed a more precise knowledge database that is based on the data culled from the pilot questionnaires. 7) To scrutinize the formation flying pattern, we can conduct a complicated simulation by using simpler rules for the helicopter flight pattern. As the result of a practice simulation, we found a way to conduct a simulation that is more complicated and realistic than any other existing model.