摘要
将模糊神经网络用于空中目标威胁估计,阐述了网络结构的确定和算法实现过程。由于模糊神经网络不需要对象的准确模型。它以分布的方式存储信息,利用网络的拓扑结构和权值分布实现非线性映射,在神经网络框架下引入模糊规则,使网络中的权值有明显的意义,且保留了神经网络的学习机制,对权值的学习采用改进BP算法。仿真结果表明了该方法的有效性,为进行有效的目标威胁评估提供了一条新的思路。
A Fuzzy Neural Network （FNN） is used to threat assessment for Aerial Target. The design of network structure and its algorithm realization are described. As a FNN can be used without the exact model for an object and it can store information in a distribution manner. A nonlinear mapping can be realized by using topo-logical structures and weights of the neural network. Fuzzy rules are introduced in the neural network which gives these weights a definite meaning and reserves the training system of the neural network. These weights are trained through a improved BP algorithm. Result shows the validity of this method. This paper presents a new method for target threat assessment.
出处
《微计算机信息》
北大核心
2007年第34期268-270,共3页
Control & Automation
基金
中科院二期创新项目(C04708Z)
关键词
模糊神经网络
威胁估计
BP算法
Fuzzy Neural Network, threat assessment, BP algorithm
