Spectrum Sensing In Cognitive Radio Technology

This thesis mainly investigates the spectrum sensing technique in cognitive radio context. As a revolutionary smart spectrum sharing technology, CR can significantly improve the spectrum utilization and draw more and more attention within these years. The core problem of constructing the actual cognitive radio networks is how to find the spectrum hole and identify the license users. So spectrum sensing, as a solution to this problem plays a vital role in CR realization.At present, the spectrum detection technology mainly includes two aspects. One is the local spectrum detection technology, that is, a single node detects the frequency occupation state based on its received signal. The other is multiple nodes sensing, that is, cooperative spectrum sensing among multiple nodes can enhance the detection sensitivity and detection accuracy and reduce the performance requirements of single node. In this paper, we focus on these two aspects.In chapter 2, an overview of the existing spectrum sensing techniques in CR is provided. At first, some local spectrum sensing schemes and cooperative spectrum sensing which belong to transmitter testing in cognitive radio network are introduced, and the advantages and disadvantages of local spectrum sensing are analyzed respectively. And then receiver testing including interference-based detection and local oscillator leakage power detection is presented and discussed.Chapter 3 focuses on the cyclostationary-based detection. Due to the distribution of carrier to the cyclostationary feature, we propose two low-complexity detection methods in which the received signal is down to IF. Cyclostationary feature of OFDM signals is explored and detection of DVB-T signals is studied. Compared to other local spectrum sensing methods, the simulation results show that cyclostationary-based detection algorithm achieves good performance.In chapter 4, this thesis investigates the cooperative spectrum sensing and its data fusion technique. According to the different type of local sensing data, the data fusion can be classified into two types: the local sensing decision fusion (also called hard combination) and the local sensing data fusion (soft combination). In hard combination, this thesis researches a softened hard combination named two-bit fusion scheme. In soft combination, this thesis studies the optimal linear fusion rule, maximal ratio combination and equal gain combination. At last, in order to improve the sensing performance which can be severely degraded when the sensing observations are forwarded to fusion center through fading channels, we apply the clustering method into cooperative spectrum sensing by forming all the secondary users into a few clusters and selecting the most favorable user in each cluster to report sensing results. This method can exploit the user selection diversity so that the sensing performance can be enhanced. We employ both the EGC soft combination scheme and two bit hard combination to improve the detection performance and reduce the transmitting overhead. Compared with conventional sensing scheme, numerical results show that the sensing performance is improved significantly.Finally, in Chapter 5, the thesis of the text is summarized, pointing out the direction for further research.