Research On Spectrum Sensing Technology Based On Free Probability Theory

Cognitive radio(CR), as an effective technology to improve the spectrum utilization and solve the spectrum scarcity problem, has received extensive attention and research. It allows the cognitive user(CU) to use the idle licensed spectrum of primary user(PU) for communication. The foundation of CR is spectrum sensing whose task is to detect whether the licensed spectrum of PU is idle or not, so its performance is critical to the CR system.Recently, free probability theory(FPT) has developed rapidly and become one of the important tools to solve the wireless communication problems. It is a main branch of random matrix theory(RMT) to portray a close relationship between two random matrices and their sum or product matrix.This thesis will integrate spectrum sensing with FPT. First, due to the poor sensing performance of the existing spectrum sensing algorithms in low signal to noise ratio(SNR) or with a small sample, a FPT-based algorithm in the time-invariable sensing channel is introduced while a new one in the time-varying sensing channel is proposed. They can extract the average received signal power from the received sample covariance matrix by setting up and solving the asymptotic freeness equation. The simulation results show that they have the fast convergence and can effectively improve the sensing performance in the case of low SNR and a small sample.Then, owing to that the data fusion rule will lead to a lot of time, energy and spectrum consumption of CU to transmit the observation data as will as the high load and slow response of fusion center(FC), a new cooperative FPT-based spectrum sensing algorithm with decision fusion rule is proposed. Here CUs only need to transmit the local decisions to FC for processing through the reporting channel. Simulation results demonstrate that the sensing performance of the proposed algorithm is satisfying and much higher than that of the RMT-based algorithm, even though the latter uses data fusion rule.Finally, in order to reduce the overhead of CR system and overcome the effect of non-ideal reporting channel on cooperative spectrum sensing, a new cooperative FPT-based spectrum sensing algorithm with the mechanisms of clustering and selection is proposed. On the one hand, the idea of clustering is introduced into the spectrum sensing and the FPT-based algorithms are used in each cluster; on the other hand, the cluster heads only send the clustering decisions which represent the PU’s signal is present to FC. Both of them can reduce the overhead and improve the sensing performance effectively. The simulation results show that the sensing performance of new algorithm is superior to that of the algorithm with the traditional OR decision fusion rule.