Abstract
Originated from military domain, Situation Awareness (SA) is proposed with the aim to obtain information superiority through information fusion and thus to achieve decision superiority. It requires collaboration of multiple agents to perform works like detection and perception of the environment, correlation of the obtained information, reasoning of the implicated meaning under explicit phenomenon, and situation assessment. Due to the highly distributed nature of data collection, and all the information processing are performed in a distributed system of systems (SoS) manner, Multi-Agent System (MAS) approach emerges great advantages for this application domain in an essential and natural way. The principal goal of this study is to exploit the semantic web related technologies and high level architecture (HLA) to construct an ontology-driven multi-agent information fusion framework to enhance situation awareness. In this dissertation, an information fusion framework that integrates ontology-based knowledge representation model, semantic enhanced agent communication infrastructure and multi-agent architecture is constructed. Ontology based approach is implemented not only for the agents’ internal knowledge representation, but also for the external communication to enhance interoperability. The agents orchestrate following JDL model to share ontology enabled domain information, thus to build consistent holistic picture about the situation. The multi-level ontology merging paradigm is followed in this work for the knowledge representation and information exchange. Firstly, Military Scenario Ontology (MSO) and Battle Management Ontology (BMO) are defined according to corresponding reputable standards from the modeling and simulation society as the domain ontology. Furthermore, we construct the Situation Awareness Ontology (SAO) as the core ontology to integrate MSO, BMO and even other publicly defined ontologies for high-level information fusion. The SAO is composed of objects, relations and events descriptions necessary for capturing information for further cognition, reasoning and decision making. Military doctrine and domain knowledge are expressed as Horn clause type rules for reasoning and inference. Multi-layered semantic information fusion that integrates SAO, semantic web technologies and rule-based reasoning can be conducted. In this research, we create a full RDF/OWL based agent memory model. The internal domain knowledge model about scenario, plans, reports are all expressed using RDF/OWL (Web Ontology Language). Agent communications are all also composed of RDF, OWL and SPARQL directly. An experimental scenario is presented to demonstrate the feasibility of this architecture. There are three contributions made in this study: (1).Ontology-driven information fusion model: we use ontology to integrate domain standards to build a situation awareness knowledge representation model including facts, relations and rules. (2). Semantic enhanced multi-agent communication infrastructure: the situation awareness knowledge representation model is used not only for agent internal knowledge management but also for communication content representation. (3). Multi-agent information fusion framework: situation awareness can be achieved after the hierarchical information flow and interaction among agents.