摘要
地声参数的不确定性对水声传播具有重要的影响。通过贝叶斯理论建立水声环境不确定性推理模型,理论推导了地声参数的似然函数以及地声参数和传播损失的后验概率密度,并采用MCMC(Markov Chain Monte Carlo)进行了仿真计算,给出了地声参数的二维后验联合概率密度和一维边缘概率密度,在此基础上对传播损失的不确定性进行了估计,得到了传播损失80%的可信区间。仿真和实验结果表明,该方法适用于地声参数反演和不确定性估计,并能获取因地声参数不确定性导致的传播损失不确定性估计。
Uncertainty of geoacoustic parameters has a great effect on acoustic propagation, tn tins paper, rne uncertainty inference model of the acoustic environment is established via Bayes theorem. Then the likelihood function of the geoacoustic parameters and the posterior probability distribution of the geoacoustic parameters and transmission loss are deduced. Furthermore, a simulation is made using Markov Chain Monte Carlo (MCMC). Two-dimensional and one-dimensional posterior probability densities of the geoacoustic parameters are given. Meanwhile, the uncertainty of transmission loss is estimated. The mean and 80% credibility interval of transmission loss in different depths are given. The results of the simulation and experiments show that the method is good at geoacoustic parameters inversion and uncertainty estimation, and it can get the uncertainty estimation of the transmission loss that is caused by uncertainty of the geoacoustic parameters.
出处
《应用声学》
CSCD
北大核心
2015年第1期71-78,共8页
Journal of Applied Acoustics
基金
中国博士后科学基金项目(20110491884)
总装预研基金项目(9140A03060213JB15039)
关键词
贝叶斯
后验概率密度
地声参数
马尔可夫链蒙特卡洛
Bayesian, Posterior probability density, Geoacoustic parameters, Markov Chain Monte Carlo