论文部分内容阅读
对观测器数目小于源信号数目的欠定盲源分离进行了研究,提出了一种基于位势函数的稀疏信号欠定盲源分离方法,该方法从能量的观点出发,通过构造位势函数,将寻找混合信号在直线方向上的聚类问题转化为寻找累积位势函数的局部极大值问题,从而准确的估计出源信号数目和混合矩阵,克服了通常的基于k-means聚类的混合矩阵估计法需预先给定源信号数目的缺点.利用仿真信号检验了该方法的有效性.基于信号频域稀疏性假设,将该方法应用于欠定条件下的滚动轴承振动故障信号的盲分离,较好地分离出了故障信号.
The underdetermined blind source separation with the number of observer less than the number of source signals is studied. A method of sparse signal underdetermined blind source separation based on potential function is proposed. From the energy point of view, this method constructs the potential function, The problem of finding the mixed signal in the linear direction is transformed into the problem of finding the local maximum of the cumulative potential function so as to estimate the number of the source signals and the mixing matrix accurately and overcome the usual hybrid based on k-means clustering The matrix estimation method needs to give the shortcoming of the number of source signals in advance.The validity of the proposed method is verified by simulation signals.Based on the signal sparseness assumption in frequency domain, this method is applied to the blind separation of vibration fault signals of rolling bearings under undetermined conditions, Better separation of the fault signal.