低信噪比下多目标调制谱轴频自动检测算法

An automatic detection algorithm for multi-target modulation spectrum shaft frequency under low signal-to-noise ratio

针对低信噪比条件下,水声目标辐射噪声中多目标调制谱轴频检测较困难的问题,提出一种联合蚁群算法与谐波库匹配算法的多目标调制谱轴频检测算法。算法利用排序截短法剔除调制谱中的趋势项,并提出一种基于蚁群算法的线谱提取算法,用于提取低信噪比下调制谱中的线谱,该算法根据线谱特征建立一种新的代价函数来替代传统蚁群算法中距离这一寻优标准,可实现低信噪比下线谱地自动提取;最后根据所提线谱建立谐波库,通过与谐波库匹配实现轴频地自动提取。仿真和海试数据验证表明,该算法在低信噪比下可以较好地提取线谱,并自动提取多目标的轴频,效果较好。

Aiming at the problem of difficult detection of multi-target modulation spectrum shaft frequency in underwater acoustic target radiated noise under the condition of low signal-to-noise ratio,a multi-target modulation spectrum shaft frequency detection algorithm combining the ant colony algorithm and the harmonic library matching algorithm was proposed.The algorithm used the sorting truncation method to eliminate the trend items in the modulation spectrum,and proposed a line spectrum extraction algorithm based on the ant colony algorithm to extract the line spectrum in the modulation spectrum under low signal-to-noise ratio.The algorithm established a new cost function based on the characteristics of the line spectrum to replace the optimization criterion of distance in the traditional ant colony algorithm,which could realize the automatic extraction of the line spectrum with low signal-to-noise ratio.Finally,a harmonic library was established according to the proposed line spectrum,and the shaft frequency was automatically extracted by matching the harmonic library.Simulation and sea trial data verification show that the algorithm can successfully extract line spectra under low signal-to-noise ratio,and automatically extract the shaft frequency of multiple targets,with satisfied results.