Building the Knowledge Base through Bayesian Network for Cognitive Wireless Networks

Tactical communication networking faces complexity, heterogeneity, and reliability requirements. The emerging research area of cognitive networks offers a potential for dealing with these problems. A key feature of cognitive networks is the knowledge base, which is produced during the process of learning and responsible for the decision making. We propose a cognitive network model integrated with the knowledge base, which is a primary part of cognitive networks. And then we focus on the construction of the knowledge base and the expression form of the knowledge in the model. In this paper, we use the Bayesian Network (BN) to construct the knowledge base, which is a unique tool for creating a representation of the dependence relationships among network protocol parameters. The data structure of the dependence relationships of the BN is translated into the knowledge which is expressed by the probability. In the simulation experiments, we create the BN through the sampling data to construct the knowledge base using the mathematical tool MATLAB and prove the efficiency of our cognitive network model for optimizing network performance in the OPENT simulation platform.