基于贝叶斯混合源分离方法的声纳探测误差模型

Sonar Detection Error Model Based on Bayesian Unobservable and Observable Source Separation

声纳的探测误差模型是声纳仿真的核心技术。针对现有的声纳误差白噪声模型仿真逼真度低的问题,提出了基于贝叶斯混合源分离方法,建立了声纳的探测误差模型,提高了声纳误差模型的性能。与传统的白噪声仿真方法相比,基于贝叶斯混合源分离方法的声纳误差仿真模型具有较好的逼真度,可以有效地对声纳探测误差进行建模和预测。

The sonar detection error model is the core technology in sonar simulation. In order to solve the problem that Ganssian noise model has low fidelity, the paper puts forward a method based on bayesian unobservable and observable source separation to eatablish the sonar detection error model in order to improve the fidelity of sonar simulation. Compared with conventional Gaussian noise model, the proposed model shows higher performanceof fidelity which could effectively mod- el and predict the sonar detection error.