A New Intelligent Data Forwarder with Ontological Agent and Concept Learning

Communication systems within MDA use a variety of message formats such as TADIL J, TADIL K, XML. To facilitate interoperable data communications, the messages need to be converted between the formats. One of the major problems faced during data conversion is data loss. Communication in the tactical network must be efficient, secure and reliable. The network must remain operational under dynamic topology. Existing routing algorithms cannot satisfy these requirements and cause data to be forwarded to wrong locations. To overcome these problems with traditional data conversion methods, we propose a novel ontology-based data conversion agent which will have an ontology defined for each of the message formats. There also exists a metadata of common vocabulary shared by all local ontologies. We further propose a novel intelligent data conversion agent which converts data between formats and also learns to extend its ontologies with new concepts and features. The concept clustering and fusion techniques will be used to classify the concept either as new or as synonymous to some existing concept. Reinforcement learning based routing method is proposed to learn the optimal routing policy to reduce data forwarding latency. Exploration methods are proposed to adapt to network topology changes.