从短波无线电通信检测在处理时间方面的实际需求出发,提出一种基于数据流趋势分析的无线电信号检测方法。分析了基于增量递推最小二乘算法构建数据流线性回归模型的方法,针对广义似然比检测法用于信号检测虚警较高的问题,提出了基于置信区间的数据流趋势分析中变化点检测方法,有效地降低了虚警。无线电检测数据动态检测实验验证该方法的有效性。
A communication signal detection algorithm based on data stream trend analysis is presented in order to meet the need of time-consuming feature.An algorithm for building data stream linear regression model is analyzed.As to the problem of false alarm when detecting the signal with the generalized likelihood ratio test, a change-point detection method is proposed based on confidence interval,which reduces the false alarm effectively.The experimental results of communication datasets demonstrate its efficiency.