全局动力学模型用于水声信号分类仿真实验研究

A Simulation Experiment of Underwater Acoustic Signal Classification Using Global Dynamical Models

水下弱信号的检测和分类是水声信号处理领域中的难题。本文研究了一种使用混沌动力学系统理论中全局动力学模型进行水声信号检测和分类的方法,用仿真数据对所建模型的有效性进行了验证。通过将模型拟合到几种舰船辐射噪声,成功地对其进行了分类。仿真结果表明,该方法在高噪声背景下(信噪比为0dB时)仍能对信号进行分类,且分类概率可以从模型系数空间的系综统计特性中直接计算得到。

Detection and classification of weak underwater signals are difficult problems in processing acoustic signals. This paper considers a method for using global models of chaotic dynamical systems theory to perform detection and classification of underwater signals, and it proved to be efficient by the simulated data. Then through fitting this model on radiation noises of several naval vessels, we successfully classified the different classes of signals. The simulation results show that classification in high noise regimes (0 dB SNR cases) can be performed, and classification probabilities can be directly computed from ensemble statistics in the model coefficient space.