Technique of Multi-Domain Local Cognition in Cognitive Networks

In cognitive networks (CN), the cognitive process that senses the multi-domain environment mainly depends on Multi-Domain Local Cognition (MDLC) as it not only establishes the multi-domain cognition information (CI) databases for the whole process but also performs the acquirement of mass CI. To the best of our knowledge, this is the first work to investigate MDLC by quantifying. To reveal the principles within MDLC, the model of Single Domain Local Cognition (SDLC) is considered firstly. Based on the model, a method named Multi-dimension Edge Detection (MED) is proposed, which can divide the parameter space describing the environment of SDLC into different areas and represent each area with an identifier. Specifically, the MED method involves four steps: acquiring parameter vector samples, classifying samples by the boundary surfaces where they located, reconstructing the boundary surfaces, and representing each area with an identifier. By applying this method in each domain of MDLC, the CI which consists of the identifiers coming from different domains can be got. The paper concludes by presenting a simple example to illustrate the feasibility of MED.