Robust signal processing algorithms for RF complexity reduction in cognitive radio systems -

Abstract

The need for an efficient spectrum usage policy is becoming more pressing due to the rapid growth of wireless communications applications and the under-utilization of most licensed bands. Cognitive Radio arises to be a novel solution to this spectrum wastage problem by allowing the dynamic allocation of unused bands. In these scenarios, a secondary or unlicensed user detects, exploits and utilizes an unoccupied spectrum band, called a white space, reserved for primary users. Thus, two major functions of a cognitive radio transceiver are scanning the spectrum for a potential white space and transmitting a suitable adaptive signal. In real-world dynamic scenarios, this highly adaptive radio technology presents unique signal processing challenges and requires specific algorithms for spectrum sensing and spectrum sculpting. In an Ultra-Wide Band (UWB) scenario, secondary users are required to scan bands that are of hundreds of megahertz in width. This puts strict hardware requirements on the RF front-end, relating to the linearity, dynamic range and sensitivity of the circuitry. On the other hand, RF impairments could dramatically degrade the performance of the secondary user. These imperfections result from the large spectrum to be sensed and from the low-cost implementations of the analog RF front-ends. In such scenarios, advanced signal processing algorithms are required not only to sense the incumbent spectrum but also to improve the radio sensitivity by mitigating the RF impairments effects to relax the hardware requirements. In this dissertation, we investigate and propose signal processing algorithms to build a highly adaptive, fully functional cognitive radio transceiver working in the UWB range. At first, we start by studying basic spectrum sensing algorithms, we then move to assess the performance of these algorithms against RF impairments. A blind algorithm is selected, improved and implemented in real-world scenarios. A mitigation stage is devised and added to the sensing scheme in order to obtain a r