Real Time Hardware Implementation of Cyclostationary Spectrum Sensing for Various Modulation Types Using USRP

Abstract

Cognitive radio (CR) is an emerging technology to solve the spectrum scarcity problem. Spectrum sensing is the most important step in CR. Spectrum sensing techniques can be classified into three main types: energy detection, matched filter, and cyclostationary. Cyclostationary sensing technique is more accurate than others since it can differentiate between signal and noise by using autocorrelation process without needing the previous knowledge about primary user (PU) signal. In this paper, we design and implement the cyclostationary feature detector in a real environment using the GNU-Radio framework and the universal software radio peripheral (USRP) N210 less complexity in the design as compared to the designs in the literature. The proposed design is tested under various types of PU signals. These types are classified into three scenarios: the first scenario is a single carrier (QPSK, 64QAM, and 256QAM), the second scenario is a multi-carrier (OFDM with 64 subcarriers, 64QAM data channel), while the third scenario is assumed that the PU is absent (only Gaussian noise). The PU signal is implemented using AWG (Arbitrary Waveform Generator) 7122c from Tektronix. The distance between PU (AWG) and secondary user (SU) (USRP N210) is 10 m, which is longer than those used in the literature. The experiment results show that the proposed system is applicable for various types of signals (single carrier, multicarrier (OFDM), and Gaussian noise only) with satisfactory performance. In the case of the presence of PU (single carrier and OFDM), the peaks of cyclostationary feature detection occur in the center and cyclic frequencies. When PU is absent (Gaussian noise only), the results show that there are no peaks at the cyclic frequencies and there is a small peak at the center frequency, which confirms the right operation of the implemented system.