Mobile agent-based distributed fusion (MADFUSION) system

Abstract

This paper considers a system architecture referred to as the mobile agent-based distributed fusion (MADFUSION) system. The system environment consists of a peer-to-peer ad-hoc network in which information may be dynamically distributed and collected via publish/subscribe functionality implemented at each node of the network to facilitate data sharing and decision making in Level 2 fusion. The Level 2 decision making process implemented in the system consists of the enhanced doctrinal template matching (EDTM) algorithm which is shown to be an improvement over the pre-existing doctrinal template matching algorithm. This algorithm is developed to operates on information obtained from lower layer fusion processes in order to identify aggregated groups of entities. The template matching algorithm is shown to be an improvement over a previously existing algorithm. The MADFUSION system is proposed to extend the client/server architecture of various publish/subscribe applications to an architecture providing decentralization, re-configurability, mobility, attainability and prevention of single points of failure. The system is implemented in a wireless ad-hoc network (802.11b) and performs the publish/subscribe functionality through the implementation of a mobile agent based framework. The software agents travel deterministically from node-to-node carrying a data payload consisting of information which may be subscribed to by users within the network. Within this system, situation awareness (Level 2 fusion) can be sought by using these multi-domain information sources (GMTI, Video, or SAR) for evaluation at each node with different distributed information fusion algorithms.