An on-line blind source separation （BSS） algorithm is presented in this paper under the assumption that sources are temporarily correlated signals. By using only some of the observed samples in a recursive calculati...An on-line blind source separation （BSS） algorithm is presented in this paper under the assumption that sources are temporarily correlated signals. By using only some of the observed samples in a recursive calculation, the whitening matrix and the rotation matrix could be approximately obtained through the measurement of only one cost function. SimNations show goad performance of the algorithm.展开更多
针对雷达接收机在现代战场复杂电磁环境下接收到的混叠信号,提出了一种基于二阶矩的信号盲源分离方法。在混合信号球化过程中,对于具有加性白噪声的模型,构造了一组新的协方差矩阵,在信噪比不是很高的情况下,使其不会影响分离结果。在...针对雷达接收机在现代战场复杂电磁环境下接收到的混叠信号,提出了一种基于二阶矩的信号盲源分离方法。在混合信号球化过程中,对于具有加性白噪声的模型,构造了一组新的协方差矩阵,在信噪比不是很高的情况下,使其不会影响分离结果。在协方差矩阵对角化过程中,采用自然梯度的方法,避免分离矩阵更新过程中的求逆问题,提高了算法的实时性。仿真实验证明,在信噪比为-10 d B的条件下,对比Fast ICA算法,所提算法分离精度高,收敛速度快,为进一步的信号识别提供可靠依据。展开更多
基金This project was supported by the National 863 project (2001AA422420 -02)
文摘An on-line blind source separation （BSS） algorithm is presented in this paper under the assumption that sources are temporarily correlated signals. By using only some of the observed samples in a recursive calculation, the whitening matrix and the rotation matrix could be approximately obtained through the measurement of only one cost function. SimNations show goad performance of the algorithm.
文摘针对雷达接收机在现代战场复杂电磁环境下接收到的混叠信号,提出了一种基于二阶矩的信号盲源分离方法。在混合信号球化过程中,对于具有加性白噪声的模型,构造了一组新的协方差矩阵,在信噪比不是很高的情况下,使其不会影响分离结果。在协方差矩阵对角化过程中,采用自然梯度的方法,避免分离矩阵更新过程中的求逆问题,提高了算法的实时性。仿真实验证明,在信噪比为-10 d B的条件下,对比Fast ICA算法,所提算法分离精度高,收敛速度快,为进一步的信号识别提供可靠依据。
