Research On Key Technique Of Spectrum Sensing In Cognitive Radio Systems

By exploiting the existing wireless spectrum opportunistically, cognitive radio technology is developed to solve current wireless network problems resulting from the limited available spectrum and the inefficiency in spectrum usage. A cognitive radio (CR) is designed to be aware of and sensitive to the changes in its surrounding, which makes spectrum sensing an important requirement for the realization of cognitive radio networks. Spectrum sensing enables CR users to detect spectrum holes without causing interference to the primary networks. This dissertation will make a deep study of the spectrum detection in wireless regional area networks.Firstly, several techniques of spectrum sensing are discussed and the concept of noise uncertainty is introduced. A spectrum sensing algorithm based on information theory in CR is considered to avoid the influence by noise uncertainty. By introducing the nonuniform quantization, the spectrum entropy-based detection scheme is proposed. The proposed scheme can improve quantization performance and maximize entropy value so as to improve the detective performance. The simulation results show that the scheme can obtain approximate 3dB gain in detection performance than spectrum entropy-based detection using uniform quantization. Furthermore, the simulation results verify the robustness against noise uncertainty, and show that the proposed scheme outperforms energy detection under noise uncertainty situation.Secondly, to the fading or hidden-station problems existed in single node detection, several hard decision methods and soft decision methods based on entropy detection in cooperative spectrum sensing are proposed in the dissertation. Fusion rules of cooperative spectrum sensing are compared by computer simulation under each node with same Signal to Noise Ratio (SNR) and different SNR. Simulation results show K out of N rule obtains the best performance under same SNR condition. And it increases detection probability by 50% than single node detection for a target false-alarm probability of 0.1 and SNR of -10dB. While maximal ratio combination (MRC) rule of the soft decision is the optimal algorithm under different SNR conditions. The simulation results show that MRC rule improves detective performance by 10% than K out of N rule for a target false-alarm probability of 0.1.