In Cognitive Radio, Spectrum Sensing Technology

Cognitive radio (CR) technique is considered as one of the solutions of current spectrum resource scarcity. The core idea of cognitive radio is to exploit the licensed but underutilized spectrum, in order to improve the spectrum efficiency. In cognitive radio systems, avoiding interference to primary users is the most important constraint. And spectrum sensing is the key technique to achieve this goal. In order to improve the performance of spectrum sensing, cooperative spectrum sensing is proposed, which can recover the limits of spectrum sensing by only one node.This thesis mainly investigates the cooperative spectrum sensing technique. According to the different research direction, it can be divided into three parts:The first part of this thesis introduces several local sensing algorithms and analyse their advantages and disadvantages respectively. As the first CR system—IEEE 802.22 wireless regional area networks (WRAN) aims at exploiting the spectrum band which the TV systems operate on, therefore, TV signal is one of the main PU signals. This thesis proposes three detection algorithms of the Digital Video Broadcasting-Terrestrial (DVB-T) signal. The simulation results show that these algorithms have good performance.In cooperative spectrum sensing, the local sensing data need transmit to the central node. In order to reduce the transmitting overhead, the local sensing data should be quantized into finite bits. The second part of this thesis investigates the quantization schemes in cooperative spectrum sensing. For the prior knowledge of the PU signal is not already known by the central node, the classical quantization schemes are not fit the cognitive radio context. Therefore, this thesis proposes two kinds of quantization schemes to solve this problem. The first one called locally optimal quantization scheme has optimal performance in the low SNR environment, and performs well in high SNR environment although it is not optimal. Another scheme is not optimal, but it does not need any prior knowledge.The last part of this thesis investigates the sensing data fusion technique of cooperative spectrum sensing. According to the different type of local sensing data, the data fusion can be classified into two types: the local sensing decision fusion and the local sensing data fusion. In the decision fusion, this thesis proposes a decision fusion scheme based on the optimal rule. In the sensing data fusion, this thesis derived the optimal fusion rule and proposed an estimation algorithm to estimate the unknown part of the optimal rule. Simulation results show that the proposed schemes perform better than the existing schemes. Meanwhile, it could approach the optimal performance if more bits'sensing data are used in the fusion.