Research On Wideband Spectrum Sensing Based On Compressive Sensing In Cognitive Radio Networks

With the rapid growth of wideband wireless communication services, the imbalance between supply and demand of spectrum resource has become increasingly conspicuous day by day. Spectrum sensing is prerequisite for the realization of the cognitive radio system. It is of great significance to increase the utilization efficiency of the spectrum and protect the licensed users free from interference.Since the traditional spectrum sensing schemes are restricted to the A/D sampling rate, so each sensing frontend can only senses within a relatively narrow frequency range, which will cause time delay and uncertainty. Therefore the wideband spectrum sensing is promptly needed. But because of cognitive radio users process the receiving signal at Nyquist sampling rate, so the wideband spectrum sensing become more difficult to ask for the radio frequency front-end sampling rate, which dramatically increases the complexity of spectrum sensing. The introduction of the compressive sensing technologies perform wideband spectrum sensing can greatly reduce the number of sample points and the data quantity processed, can also shorten the detection time and lower the requirements of hardware.Taking account of the impact of the low detection probability in a low SNR and the uncertainty of compressive coefficient, a modified adaptive spectrum sensing algorithm based on compressive sensing is proposed. The optimal compressive coefficient is selected adaptively according to SNR to obtain the primary users’ information. When SNR is lower, the bigger compressive coefficient is selected to ensure the spectrum sensing accuracy. When SNR is higher, the algorithm adaptively reduces compressive coefficient to shorten the execution time dramatically without the influence of the detection accuracy, which can improve the efficiency of the spectrum sensing. The scheme uses the basis pursuit algorithm. Since the power spectrum density estimation takes priority over signal reconstruction at far low Nyquist sampling rate, without loss of precision the algorithm realizes fast and reliably spectrum sensing. In this study the simulation results show that the algorithm can reduces the uncertain affect of compressive coefficient resulting in achieving a reasonable compromise between accuracy and the cost of detection at low SNR and fading channel.However, the disadvantageous factors in the complex variable wireless propagation environment, such as fading, shadowing and multipath fading and so on, degrade the spectrum sensing performance for single cognitive radio user seriously. To solve this problem, we propose a cluster-based cooperative compressed spectrum sensing algorithm. By utilizing the users’ spatial diversity, within cluster fusion decision method based on the decision fusion rule and the energy fusion rule, respectively, the algorithm improves overall spectrum sensing performance for cognitive radio networks. Simulation results indicate that the scheme is significantly better than traditional compressed spectrum sensing scheme, which can perform spectrum sensing effectively at the condition of low SNR and sub-Nyquist sampling rate and then reduces the performance impact of cooperative spectrum sensing over imperfect wireless fading reporting channel.