摘要
针对同频线性调制混合信号单通道盲分离中的时延估计问题,提出了一种基于自相关函数的估计方法。该方法利用接收基带信号的自相关函数建立关于时延的非线性方程,并通过延迟采样进一步构造适定方程组,然后利用梯度下降算法对方程组进行迭代求解,完成整个估计过程。理论分析表明,时延估计的误差正比于自相关函数估计的误差,随数据量增大不断减小。仿真结果验证了该方法的有效性。
An auto-correlation function based method is presented for time delay estimation in single-channel blind separation of co-frequency linear modulated signals. The method establishes equations on the time delays using the auto-correlation functions of the received baseband mixture, constructs an overdetermined equation array by sampling at different time instants, and then finishes the whole estimation by solving the equation array iteratively using a gradient descent algorithm. The mathematical analysis shows that the estimation error is directly pro- portional to the estimation error of the auto-correlation functions, thus reducing continuously with increasing data size. The validity of the proposed method is verified by computer simulations.
出处
《数据采集与处理》
CSCD
北大核心
2010年第4期449-453,共5页
Journal of Data Acquisition and Processing
关键词
混合信号
线性调制
单通道
时延估计
mixture signal
linear modulation
single-channel
time delay estimation