动态匹配追踪中利用连续相位求取瞬时频率

动态匹配追踪算法将信号的瞬时特征作为先验信息引入到信号分解中,缩小了时频原子库的搜索范围,提高了匹配追踪效率。但在利用瞬时相位获取瞬时频率信息时会得到无物理意义的负频率,不能直接应用于匹配追踪,已有文献并没有给出瞬时频率存在负异常时的解决方案。基于此,笔者通过研究主值相位与解缠相位之间的差异,提出了利用连续相位求取瞬时频率,并推导了连续相位的求取公式;然后通过实例分析,获得了反映三种不同相位数值变化的趋势图;最后通过对比三种相位求取的瞬时频率,证明由连续相位求取的瞬时频率不存在负异常,可直接应用于匹配追踪算法。

The calculation of the instantaneous frequency using the continuous phase in dynamic matching pursuit algorithm

The instantaneous characteristics of signals, as the priori information, have been introduced into the signal decomposition which can reduce the scanning range of the atomic library and improve the efficiency of the dynamic matching pursuit algorithm.Howev?er,when we use the instantaneous phase to obtain the instantaneous frequency,there will be negative frequency without physical mean?ing and the negative instantaneous frequency can't be applied to matching pursuit.As far as we know,there is no effective solution to o?vercome the negative instantaneous frequency in matching pursuit. In view of such a situation,the authors,through studying the differ?ence between the principal phase and the unwrapping phase,propose a method to obtain the instantaneous frequency by introducing the continuous phase.Then the formula of calculating continuous phase can be deduced.In the numerical data tests,the authors present the trend graphs that reflect the changes of the three different phases.Finally,the comparison of the instantaneous frequencies obtained by the three phases proves that the instantaneous frequency obtained by the continuous phase avoids the negative anomalies,and hence it is a good choice for matching pursuit algorithm and can be used directly.