Abstract: Energy Detection（ED） is one of the most common method of signal detection, which has the advantage of low complexity and unrequired prior information of signals. When the noise variance is known, the energy detection algorithm can obtain better detection performance. In most cases, the variance of the noise is required to be estimated, so the uncertainty of the noise variance estimation will have a big impact on the performance of the energy detection algorithm. To minimize the effects caused by the uncertainty of noise variance, an improved multi-antenna energy detection scheme is proposed. The program takes a simple combination of the signal with multiple antennas to construct a decision statistic, which is independent of the noise variance. Theoretical formula shows the probability of detection and probability of false alarm in this program is independent of the noise variance. Simulation results show that when the number of the antenna is more than 3, the detection performance of the program better than conventional energy detection performance.

Key words: cognitive radio, signal detection, energy detection, multi-antenna

摘要： 能量检测（ED）方法是最常用的信号检测方法之一，其具有实现复杂度低和不需要信号先验信息的优点。当噪声方差已知时，能量检测算法可以获得较好的检测性能。在大多数情况下，噪声的方差是需要预估的，因此噪声方差估计的不确定性会对能量检测算法性能造成较大的影响。为了减小由噪声方差不确定性造成的影响，提出了一种改进的多天线能量检测方案。方案将多天线采集的信号进行简单组合，从而构建出一种与噪声方差无关的判决统计量。理论公式表明，方案的检测概率和虚警概率是与噪声方差无关的。仿真结果显示，当天线的数量大于3时，方案的检测性能优于传统的能量检测性能。

关键词: 认知无线电, 信号检测, 能量检测, 多天线