The Study On Spectrum Sensing In Cognitive Radios

To solve the conflict between spectrum scarcity and spectrum under-utilization, policy makers and regulators are working on new dynamic spectrum assignment strategies. Specifically, Cognitive radio (CR) is considered as an effective way to tackle the conflict. Cognitive radio is a technology which is capable of detecting the idle spectrum and makes use of the unutilized spectrum to communicate. In Cognitive Radio, in contrast with static spectrum assignment method, other users except Primary Users (PU), named Secondary Users (SU) are permitted to make use of the unutilized channel allocated to PU. To access the unutilized channel, Cognitive radio is designed to be aware of the radio frequency environment. One of the most important components of cognitive radio technology is spectrum sensing. By sensing spectrum, SU is able to dynamically access the licensed channel according to the state of channel. Therefore, it is necessary to study on spectrum sensing.In OFDM-based systems, spectrum sensing algorithms can be divided into two groups. The algorithms in the first group utilize the cyclic prefix (CP) to perform sensing. With the help of the information about time synchronization, it achieves a good performance. Otherwise it will have a worse performance. The algorithms in the second group exploit the pilots inserted into the OFDM symbols. Compared with the first group, the latter can achieve a better performance.The main work of this thesis is as follows. Firstly, we introduce CR, energy detection and waveform-based detection, analyze the energy detection in details and make a derivation for waveform-based detection. In order to solve the problem of fading and noise in energy detection, we provide some methods to deal with it. Secondly, in order to improve the performance of CP sensing method, we analyze the distribution property of CP sensing method, and then propose a new spectrum sensing method named silding window sensing method. Simulation results show that the silding window sensing method can achieve a better performance compared with CP sensing method. Finally, in order to improve the performance of spectrum sensing algorithm based on pilots, we exploit the structure and periodicity of pilot tones to sense and a new spectrum sensing method based on pilots is proposed. Simulation results show that the proposed algorithm can achieve a better performance compared with CP sensing method.