单通道时频重叠高斯调幅通信信号盲分离方法

为了解决单通道多个时频重叠高斯调幅通信信号共信道的盲分离问题,提出了一种混合遗传与最小值搜索的盲分离算法。根据信号的特征,建立信号模型,利用接收到的一组混合信号的数据,利用混合遗传与最小值搜索的盲分离算法对所有待估计参数进行联合搜索,实现了多个信号分量的重构和分离。仿真结果表明,与传统的最小值搜索算法(初值选取误差控制在远离期望值的5%以内)和遗传算法相比,该算法有较高的数值精度和不受初值选取影响的优点。

Blind Separation Method of Single Channel Time-frequencyOverlapped Gauss Amplitude Modulation Signals

In order to solve blind separation problem of the signal-channel and time-frequency overlapped gauss amplitude modulation signals, a hybrid genetic-minimum search algorithm is proposed in this paper. At first, according to the characteristics of signals, a signal model is set up. The algorithm makes jointly search of all estimated parameters with the received a set of mixed signal data .Finally, the reconstruction and separation of the multiple signal component is realized. The simulation results show that compared with the traditional minimum search algorithm (initial value selection error controlled within 5% from expectations) and genetic algorithm, the algorithm has high accuracy and is not affected by initial value selection.