基于匹配追踪算法的自适应时频分解改进及其应用

针对匹配追踪算法计算量太大,在实际中运算时间太长的问题,结合匹配追踪的特点,采用遗传算法来搜索每次匹配的最佳时频原子,并做出改进以提高其运算的度和局部搜索能力。对信号进行分解后,引入维格纳分布,消除信号在时间-频率平面上能量密度分布的交叉项,从而充分揭示信号的时频信息。

Application of Adaptive Time-Frequency Decomposition with Modified Matching Pursuit

Computing the optimal expansion of a signal in a redundant dictionary of wave forms by using Matching Pursuit is time-consuming.So it is hard to be adopted in practice.In this paper,the optimal time-frequency atom at every pursuit is searched by using Genetic algorithm.And some improvements are provided for reducing the operation complex and boosting the ability of local search.After the decomposition the energy density distribution is described with Wigner-Ville distribution.The single time-frequency atom is drawn on the time frequency plane with its strength decided by the decomposition coefficient.In such a way the interference terms is eliminated.So the time-frequency information in the signal can be revealed sufficiently.