| Software Defined Radios (SDRs) lend themselves to flexibility and extensibility because they depend on software to implement radio functionality. Cognitive Engines (CEs) introduce intelligence to radio by monitoring radio performance through a set of meters and configuring the underlying radio design by modifying its knobs. In Cognitive Radio (CR) applications, CEs intelligently monitor radio performance and reconfigure them to meet it application and RF channel needs. While the issue of introducing computational knobs and meters is mentioned in literature, there has been little work on the practical issues involved in introducing such computational radio controls.;This dissertation decomposes the radio definition to reactive models for the CE domain and real-time, or dataflow models, for the SDR domain. By allowing such design space decomposition, CEs are able to define implementation independent radio graphs and rely on a model transformation layer to transform reactive radio models to real-time radio models for implementation. The definition of knobs and meters in the CE domain is based on properties of the dataflow models used in implementing SDRs. A framework for developing this work is presented, and proof of concept radio applications are discussed to demonstrate how CEs can gain insight into computational aspects of their radio implementation during their reconfiguration decision process. |
