Cognitive Radio Formulation and Implementation

Abstract

This paper approaches cognition on the physical and MAC layers by defining a common language of "knobs" and "meters" to discuss adaptation and learning. Cognitive radio merges artificial intelligence and software defined radios (SDR). It requires a simple language for communicating between these two levels. We define a method for doing this. We also discuss a genetic algorithm approach to perform intelligent radio adaptation, using the GNU radio platform as an example. We provide both conceptual and practical implementation details of a cognitive radio acting at the physical and MAC layers. Results presented show the promise for the genetic algorithm adaptation within the multi-objective optimization environment of the cognitive radio