Design Space Decomposition for Cognitive and Software Defined Radios

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.