Paulo Marques, COGEU
Proving that limited radio spectrum resources can be used successfully by both primary and secondary users in an economically feasible way.
Creating a flexible, live testing platform by using NI LabVIEW software and USRP™ (Universal Software Radio Peripheral) hardware to develop a cross-platform cognitive radio demonstrator that combines spectrum sensing and a geo-location database to show effective spectrum reuse.
As part of the European Commission’s Seventh Framework Program, researchers from eight European Union countries formed the COGEU Project, which focuses on cognitive radio systems to efficiently share TV white spaces (TVWS) in Europe. Ultimately, spectrum reuse promises to reduce the cost of wireless connectivity and meet the growing demands of high-data-rate wireless devices. The COGEU team is actively moving the concept of spectrum reuse from idea, through simulation, to real-world prototype, with the goal of influencing economically feasible public policy.
In the case of TV spectrum, there are two types of primary user transmissions including broadcast TV and public making and special events (PMSE) devices. The location of the broadcast towers is known and encompasses a given geographic area around a TV transmission tower (multiplexer), known as the exclusion area. In cooperation with German broadcasters, COGEU developed and populated a TVWS geo-location database for the Munich area, which applies realistic protection criteria for commercial digital TV receivers. The geo-location database stores the list of vacant TV channels for each geo-pixel; however, many PMSE applications, such as wireless microphones, occupy these TV channels and are unpredictable in time, frequency and location. This behaviour undermines the practical exploitation of valuable spectrum opportunities in TV bands for other applications. Because there is not a PMSE system with a standardized waveform, blind detection techniques are needed to determine the existence of a PMSE device without any prior information regarding its signal.
The COGEU platform demonstrator cross checks the TVWS geo-location database information with local PMSE sensing. The researchers then determined through simulation that two methods, covariance absolute value (CAV) and blindly combined energy detection (BCED), offer superior performance against other forms of PMSE detection. The PMSE sensing algorithms were integrated into the COGEU prototype. Trials in realistic scenarios showed CAV and BCED algorithms outperform the standard energy detection (ED) algorithm to evaluate gains from the more advanced detection algorithms. The proposed methods can sense a PMSE signal with a signal-to-noise ratio equal to -17 dB in a Rayleigh channel, considering a 100 ms sensing time, with 90 percent probability of detection and 10 percent probability of a false alarm.
We took a graphical system design approach to this project and used LabVIEW system design software to integrate a GPS receiver for geo-location, an NI USRP device for cognitive radio operation, Internet connectivity to the geo-location database, and a GUI. Seamless integration between the NI USRP devices and LabVIEW software, in addition to creating an intuitive UI showcasing this research, helped other engineers and policy makers to have a more comprehensive understanding of the project.
Until now, the implementation of sensing techniques for cognitive radio based on USRP software defined radios was mainly done using GNU Radio open-source software. The COGEU system requires integrating different technologies and protocols, such as online access to the TVWS geo-location database, which makes LabVIEW an ideal platform for the prototyping and proof of concept demonstration. We used many LabVIEW features for this project, including mathematics, signal processing, the NI USRP software driver, and connectivity and data communications tools.
The sensing platform relies on the NI USRP, a GPS receiver and a host PC with LabVIEW installed. The NI USRP-2920 configuration provides tuneable RF from 50 MHz to 2.2 GHz. The host PC is a laptop with a Windows OS, Gigabit Ethernet connection to link to the NI USRP radio, and a wireless connection for Internet access. The GPS device connects to the host PC by Bluetooth. Commercial tuneable FM wireless microphones were also used.
LabVIEW and the NI USRP hardware platform were key components of this research project. This platform helped the team rapidly prototype and successfully deploy the first test bed of this kind. Moving forward, the demonstration will be updated with more advanced functionality such as detection of multiple PMSE devices in a single TV channel, automatic spectrum shaping, and spectrum aggregation techniques to efficiently use vacant TV channels with protection of TV and PMSE users.
Reference papers:
Rogério Dionísio, José Ribeiro, Jorge Ribeiro, Paulo Marques, Jonathan Rodriguez, “Cross-platform Demonstrator Combining Spectrum Sensing and a Geo-location Database,” Future Networks and Mobile Summit, 4 - 6 July 2012, Berlin, Germany (to appear in the conference proceedings)
COGEU geo-location database available online:
Paulo Marques
COGEU
pmarques@av.it.pt