论文部分内容阅读
本文提出了一种改进由时频不相交分量组成信号的双线性时频分布的分辨率和可读性的方法。用修正的Xie-Beni聚类有效性指标对熵调整模糊c-均值聚类算法进行拓展将模糊聚类与密度估计相结合,实现了信号时频分量的识别和建模;信号的时频能量混合模型给出了信号分量的数目及其在时频面上所占据的区域。这些信息可以用于分离信号分量,设计适合于每个分离分量的光滑核。仿真结果表明,对于由时频不相交分量组成的信号,本方法可以识别出其中的信号分量,并得到较优的时频分布。
This paper presents a method to improve the resolution and readability of the bilinear time-frequency distribution of signals composed of time-frequency non-intersecting components. With the modified Xie-Beni cluster validity index, the entropy adjustment fuzzy c-means clustering algorithm is extended to combine the fuzzy clustering and density estimation to realize the identification and modeling of the signal time-frequency components; the time-frequency energy The mixed model gives the number of signal components and their area occupied by the time-frequency plane. This information can be used to separate the signal components and to design a smooth kernel for each separate component. The simulation results show that the proposed method can identify the signal components and obtain a better time-frequency distribution.