改进的复值非高斯最大化盲源分离算法

复值非高斯最大化盲源分离算法的收敛速度,受其采用的反双曲正弦函数计算复杂度的影响较大,针对这一问题,提出了一种收敛速度更快的改进型复值非高斯最大化算法。该算法采用一个低计算复杂度的非线性函数替代原算法中的反双曲正弦函数。该函数的计算复杂度是反双曲正弦函数的三分之一。在相同条件下,基于该非线性函数的复值非高斯最大化盲源分离算法的收敛速度相比原算法提高了1倍左右,更适合于对接收信号进行实时处理,而且信号分离误差小,同时该算法同样适用于任意非高斯的常态和非常态信源。仿真试验验证了算法的有效性。

Improved complex maximization of nongaussianity blind source separation algorithm

Due to higher computational complexity of anthyperbolic function, that was used in complex maximization of nongaussianity algorithm, influenced convergence speed of algorithm, the fast complex maximization of nongaussi-anity algorithm that based a new lower computational complexity nonlinear function was proposed. Computational complexity of the new function is about third of anthyperbolic function. In the same condition, the convergence speed of fast maximization of nongaussianity algorithm that based on the new nonlinear function was half of maximization of non-gaussianity algorithm, so it is more suitable for dealing the received signals in time. The algorithm had smaller error and could apply to any proper and improper sources. Simulation proved efficiency and correctness of the algorithm.