Abstract

In this paper, modified correlation detector based spectrum sensing technique is proposed assuming additive Laplacian noise. In the proposed modified detector, we consider a test statistic where the received signal at the cognitive terminal is correlated with primary user signal. Binary phase shift keying primary user signal is considered for simplicity. Then the received signal is raised to an arbitrary exponent P whose value ranges from 0 to 2 (0 < P < 2). Thus, the proposed detector behaves as a non-linear detector at all values of P except P = 1. At P = 1, the detector behaves as a general correlation detector or a matched filter detector. Considering the proposed test statistic, the analytical expression of detection probability and false alarm probability is derived. Performance of the proposed test statistic is presented in the form of receiver operating characteristic and detection probability vs. average signal-to-noise ratio (SNR). Optimum value of P for different values of average SNR is also obtained using simulations. The analytical expressions are validated by comparing the results with the simulation results. It is observed that the performance improves with decreasing the value of P in the proposed test statistics. Further, it is also observed that the proposed modified detector shows improved performance than conventional matched filter detector for P < 1.

中文翻译：

---

认知无线电中拉普拉斯噪声的基于修正相关检测器的频谱感知

摘要

在本文中，假设加性拉普拉斯噪声，提出了基于改进的相关检测器的频谱感知技术。在提出的改进检测器中，我们考虑了一个测试统计量，其中认知终端接收到的信号与主用户信号相关。为简单起见，考虑二进制相移键控主用户信号。然后将接收到的信号提升到任意指数P，其值的范围从 0 到 2 (0 < P < 2)。因此，所提出检测器表现为在的所有值的非线性检测器P除了P = 1。在P= 1，检测器表现为通用相关检测器或匹配滤波器检测器。考虑到所提出的检验统计量，推导出检测概率和虚警概率的解析表达式。建议的测试统计的性能以接收机操作特性和检测概率与平均信噪比 (SNR) 的形式呈现。还使用模拟获得了不同平均 SNR值的P最佳值。通过将结果与模拟结果进行比较来验证解析表达式。观察到性能随着P值的降低而提高在建议的测试统计中。此外，还观察到，当P < 1 时，所提出的改进检测器比传统的匹配滤波器检测器表现出更好的性能。