论文部分内容阅读
针对目前大多数独立分量分析(independent component analysis,ICA)算法无噪或者弱噪声假设的局限性,提出一种适用于低信噪比情况的有噪独立分量分析算法。该算法以分离信号的负熵为目标函数,采用高斯分布密度模型作为非线性函数来估计负熵,并建立了模型参数的确定准则,能够较好地抑制低信噪比下噪声的影响,最后采用人工蜂群算法对混合矩阵进行全局寻优。仿真结果表明,与其他算法相比,提出的算法可以更为精确地估计混合矩阵,能够较好地解决低信噪比下的有噪ICA问题。
Aiming at the limitation of noisy or weak noise assumption of most independent component analysis (ICA) algorithms, a noisy independent component analysis (ICA) algorithm is proposed for low signal to noise ratio. In this algorithm, the negative entropy of the signal is taken as the objective function, the Gaussian distribution density model is used as a non-linear function to estimate the negative entropy, and the determination criterion of the model parameters is established, which can better suppress the influence of noise under low signal-to-noise ratio Artificial bee colony algorithm for global optimization of the hybrid matrix. The simulation results show that compared with other algorithms, the proposed algorithm can estimate the hybrid matrix more accurately and can solve the noisy ICA problem with low SNR.