基于自适应组合窗函数的SPWVD算法及应用研究

Research on SPWVD Algorithm and its Application Based onAdaptive Combined Window Function

针对平滑伪Wigller-Ville分布(Smoothing Pseudo-Winger-Ville distribution,SPWVD)在处理不同信号时,需要采用对应的平滑窗参数才能达到最优时频分布的问题,提出了一种基于自适应组合窗函数的SPWVD算法。该算法将Rényi熵作为时频评价指标,以组合窗函数的窗长和系数作为参数库,提出将随机梯度下降和扰动随机梯度下降分别用于窗长和系数的选取,提升计算速度分别为遍历法的20倍和9倍,最终得到信号的最优时频图。将上述的改进方法应用于碳纤维复合材料的缺陷检测中,结果表明,自适应组合窗函数的SPWVD可以有效地对不同信号选取最优的窗函数参数,对比固定窗函数的SPWVD,自适应组合窗函数的SPWVD具有更好的时频聚集效果,更加完备地对信号进行了表征,并对缺陷进行了定量分析,具备缺陷检测的良好应用前景。

Aiming at the problem that smoothing pseudo-Wigller-Ville distribution(SPWVD)needs to use corresponding smoothing window parameters to achieve the optimal time-frequency distribution when dealing with different signals,a SPWVD al⁃gorithm based on adaptive combined window function is proposed.In this algorithm,Rényi entropy is taken as the time-frequency evaluation index,and the window length and coefficient of the combined window function are taken as the parameter library.Ran⁃dom gradient descent and disturbance random gradient descent are used to select the window length and coefficient respectively,and the calculation speed is 20 times and 9 times higher than that of the ergodic method.Respectively,finally the optimal time-fre⁃quency diagram of the signal is obtained.The improved method mentioned above is applied to the defect detection of carbon fiber re⁃inforced polymer.The results show that the SPWVD of adaptive combined window function can effectively select the optimal window function parameter for different signals.Compared with the SPWVD of fixed window function,the SPWVD of adaptive combined window function has better time-frequency aggregation effect,more complete signal characterization and quantitative analysis of de⁃fects.It has a good application prospect of defect detection.