神经网络在声纳建模中的应用  

Application of the Neural Network in Modeling Sonar

作　　者：白晶[1] 浦铁成[1] 

机构地区：[1]北华大学电气信息工程学院

出　　处：《北华大学学报（自然科学版）》2008年第4期374-378,共5页Journal of Beihua University（Natural Science）

摘　　要：基于声纳方程，同时考虑目标声特性和海洋环境参数，建立了主动式和被动式的声纳方程．利用海洋环境数据训练神经网络，实时解算不同环境参数时的噪声级．通过MATLAB仿真，比较径向基神经网络（RBF网络）和误差逆传播神经网络（BP网络）在模型中的作用，证明了RBF网络比BP网络能更好地满足声纳建模的需求．The performance of sonar is seriously influenced by its working conditions. There is not the clear mathematic model to describe the relations. This is the key of the sonar modeling. Through searching the influence between sonar noise level and ocean environment parameter in the sonar working conditions, a neural network method was proposed for the noise level model in the sonar modeling. Comparing the effect of the RBF neural network and the BP neural network in the sonar modeling by using MATLAB, it shows the RBF neural network can better satisfy the system requirement than the BP Neural network.

关 键 词：RBF BP 建模 仿真 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]