Research On Compressed Spectrum Sensing Method In Cognitive Radio

Spectrum sensing is one of the key technology of cognitive radio(CR) system, among which energy detection algorithm becomes the hot-spot because of its simplicity and practicality. At the same time, compressed spectrum sensing recently becomes the hot topic because of its low sampling rate and good performance. This paper firstly analyzes characteristics of those methods used both in the spectrum sensing technologies and compressed sensing technologies. On the basis of these analyses, compression spectrum perception technology model and performance improvements are mainly studied.The performance of current energy detection algorithms by only using single threshold will worsen with the decrease of SNR. To solving this problem, this paper proposes a new iterative spectrum sensing method(FACDR, False Alarm-Constant Detection Rate) which uses a double decision threshold. FACDR combines CFAR(Constant False-Alarm Rate) criterion and CDR(Constant Detection Rate) criterion to set the decision threshold. The simulation results show that perception results are close to the desired false alarm probability when reaching a certain probability of detection. Compared with only using traditional CFAR criterion and CDR criterion to set the decision threshold, the performance of perception results by using FACDR is improved greatly.In the existing compressed spectrum perception reconstruction methods, most of the traditional measuring method adopt point measurement to estimate support set. However, studies show that recognition mechanism based on point measurement has lack of robustness which could cause error. In order to solve this problem, this paper proposes a novel collaboration compression spectrum sensing method which combines energy detection and group sparse feature. This method is called NGOMP(Energy-Group Orthogonal Matching Pursuit). Cognitive users make full use of the sparse feature of the group energy signals to estimate support set. NGOMP reduces the influence on the final collaborative perception judgment caused by poor detection performance. So as to improve the performance of the spectrum detection results. Finally, experiments show that NGOMP can greatly improve the detection performance on the basis of lower sampling rates.