基于自适应分割算法的多跳频信号盲检测

传统的观点大都将跳频信号盲检测问题视为能量域的门限阈值问题，而从统计域来看，实际接收到的跳频信号是在一些未知时刻突变而在这些时刻之间保持统计平稳性的分段平稳随机信号，那么基于非平稳时间序列的各种突变检测算法就可以引入其中。分析了当前跳频突变通信信号的统计特性，给出了其高阶分段平稳的模型。将Bemaola-Galan（BG）提出的自适应分割算法推导到高阶，并将其成功应用于多个跳频突发信号盲检测和自适应提取中。仿真结果表明，该算法不需要任何先验信息，能够有效检测和提取多个突发通信信号，且性能优于传统的能量检测法。

Blind detection of multiple FH signals based on modified BG algorithm

The blind detection of frequency-hopping(FH) signals usually was treated as an energy threshold problem.However,non-stationary time series anomaly detection models will be applied when piecewise stationary is a prominent feature of real data that can be associated with regimes(segments) of different statistical properties.This paper analyzes the statistical properties of FH communication signals and suggests a high order piecewise stationary model.It derives a segmentation algorithm by Bemaola-Galan(BG) from the mean,standard deviation to the higher order.The modified BG algorithm is successfully used for blind detection and segmentation of multiple FH signals.Simulation results demonstrate the effectiveness and superiority of the proposed algorithm.