Spectrum sensing through software defined radio

Dissertação apresentada para obtenção do Grau de Mestre em Engenharia Electrotécnica e de Computadores, pela Universidade Nova de Lisboa, Faculdade de Ciências e TecnologiaA change in paradigm when it comes to controlling radio transmissions is in course. Tasks usually executed in an exclusive class of hardware systems are increasingly controlled by software systems. A deep change to the software domain is foreseeable, creating a true Software Defined Radio. At the same time this change occurs, the radioelectric spectrum is almost completely licensed. However, the spectrum is rarely used to its full extent over time, enabling its opportunistic use while the licensed devices do not communicate. This is a part of the notion of Cognitive Radio, a new kind of radio capable of using the spectrum in an opportunistic way. These two new paradigms in radio access can be combined to produce a exible and reliable radio, overcoming the issues with radioelectric spectrum scarcity. This dissertation starts an exploration in this area by combining these two paradigms through the use of an Energy Detector implemented in a Universal Software Radio Peripheral device and using the GNURadio suite. The performance of such a system is tested by calculating the Probabilities of Detection and False Alarm in real scenarios and comparing them to the expected theoretical values. A method for defining thresholds for narrowband signals is also tested based on works in Information Theory concepts, i.e.,the Akaike Information Criteria and the Minimum Description Length. The results are tested for a real transmission using two USRP platforms communicating with each other,one acting as the licensed user and the other acting as the secondary, opportunistic user. Finally, we highlight the technological work developed in this dissertation, which may support future research works through the use of the developed scripts, allowing a faster method to test algorithms with different parameterization