An agent-based data filtering mechanism for high level architecture

Abstract

In the Runtime Infrastructure (RTI) of the High Level Architecture (HLA), the purpose of the Data Distribution Management (DDM) services is to reduce the amount of irrelevant data communicated between federates (simulations) and cut network communication cost. The current DDM schemes proposed for the RTI, i.e., region-based and grid-based DDM, are both oriented to send as little irrelevant data to subscribers as possible, but they only manage to filter part of this information and some irrelevant data is still being sent. In this paper, we employ intelligent agents to perform data filtering in HLA to address these drawbacks. In our method, each time a federate subscribes to some data, intelligent mobile agents are launched to the publishers of those data. These agents will fetch the updated data, perform data filtering and then send the subscribers the exact information they require. At the same time, agents will be kept notified of the updated filtering range by keeping contact with the subscribers. Based on this idea, we implement an agent based RTI (ARTI) using the FDK RTI-kit. The RTI with region-based DDM and the RTI with grid-based DDM are also implemented to aid in the comparison. The ARTI is tested and compared with the other filtering mechanisms using the AWACS sensing aircraft simulation scenario and air traffic control simulation scenario. Experimental results show that the agent-based filtering approach performs best in data filtering and communicates only relevant data, minimizing network communication. It is also comparable in terms of time efficiency.