The Spectrum Sensing Technology Based On Multi-antenna In Cognitive Radio Systems

Nowadays, the radio spectrum was allocated for use primarily by the state, but in practice many assigned bands are underutilized in most cases, and this made the wireless resources depleting. In recent years, the cognitive radio technology provided a new way to improve the spectrum efficiency. It can continuously perceive the changes of the external environment through learning, and adapt to this change by adjusting its own internal communication mechanism. And this technology created a new situation for solving the spectrum shortage and achieving dynamic spectrum management. The spectrum sensing technology was one of the first and foremost technologies in cognitive radio.This paper mainly studies the spectrum sensing algorithm in single-user multi-antenna case. First of all, the paper reviewed the cognitive radio systems and spectrum sensing techniques, described its theoretical background and development, and then described the basic principles of diversity in MIMO and the specific calculation process of the three kinds of combined methods. Then, we presented a detailed derivation of the energy detection algorithm for high SNR and cyclostationary feature detection algorithm for low SNR, and separately compared their detection performance in AWGN channel, Rayleigh fading channel and Rice fading channel. In order to resist the adverse effects to the system detection performance by channel fading, we bring the MIMO diversity techniques into the sensing algorithm, simulated and analyzed the detection performance when using multiple antennas in case of each branch suffered the same decline. When the branches suffered different decline, we separately apply Select Combing (SC), Equal Gain Combing (EGC) and Maximum Ratio Combing (MRC) algorithm to analyze the detection performance. For problem of the high computational complexity of the total detection statistics in the cyclostationary feature detection algorithm under multi-antenna environment, the paper lastly proposed a new method, in which the signal copy in each branch was combined into a general signal, and then used it to calculate the detection statistics. The simulation results show that, this method not only reduced the computational complexity but also ensured the system’s detection performance.