Large-scale Hybrid Testbed for Cognitive Radio Networks by Real-and-Virtual Execution

Cognitive-radio networks are intended to be future spectrum-efficient systems that can adapt to their operating environments by extending software defined radio (SDR) technologies. A testbed system for cognitive radio needs to be able to deal with sensitive radio components to verify low-level protocols and functions on CR terminals, and to evaluate high-level network protocols and throughputs using various applications running on a large wireless network. We propose a testbed for cognitve-radio networks under which terminals communicate using both real CR devices and those virtually configured in the sensing environment. All the stacks from MAC/PHY to the application layer are programmable and conform to the common interface of the simulation/emulation environment for CR networks, called CWC-EF. The physical layer is either configured as a real CR device or as a virtual module with a cognitive component either by sensing real spectra or time-driven radio-feature extracted data that are generated from radio data. A real-device approach is used to verify the radio configuration and a virtual configuration is used to evaluate the network as many CR terminals are expected to be utilized. The testbed is flexible enough to re-build CR functions and protocols under any spectrum-sensing environment. CWC-EF is also intended to be used not only by local users but also remote users via the Internet.