A Zynq-based Cluster Cognitive Radio

Abstract

Traditional hardware radios provide very rigid solutions to radio problems. Intelligent software defined radios, also known as cognitive radios, provide flexibility and agility compared to hardware radio systems. Cognitive radios are well suited for radio applications in a changing radio frequency environment, such as dynamic spectrum access. In this thesis, a cognitive radio is demonstrated where the system self reconfigures to demodulate a detected waveform. The GNU Radio framework is used to provide basic software defined radio building blocks and is supplemented with FPGA accelerators. The use of GNU Radio compliant hardware interfaces allows for seamless hardware/software radio deployments. Dynamic resource mapping allows radio designers to operate at a layer of abstraction above the physical radio implementation. By establishing lower level abstraction layers, future researchers can focus on larger picture concepts such as learning algorithms and behavioral models for the cognitive engine.