基于粗糙集和ART2 神经网络的多传感器信息融合

神经网络是信息融合的一种重要方法,粗糙集理论是处理不完备信息的一种技术。本文提出了一种基于粗糙集和ART2神经网络的多传感器信息融合方法。ART2网络是一种无监督神经网络,能够实现对输入的任何模式信号自动识别和分类。而对信息融合中常遇到的数据超载问题,提出采用粗糙集与神经网络结合的方法解决。文章给出了基于粗糙集理论的组合神经网络的模型结构,最后用一个脱机手写体数字识别的实例验证了该方法的有效性。

Multi-sensor information fusion based on rough set theory and ART2 neural network

The neural network is an important way of data fusion. The rough set theory is a kind of technology to deal with incomplete information. In this paper, a method of multi-sensor information fusion based on rough set theory and ART2 neural network is presented.ART2 network is a sort of unsupervised neural network. It can achieve automatic identification and classification towards any input pattern signal. Aiming at the problem of data overloading in data fusion ,we combine the rough set theory and neural network. The paper offers the model structure of combination neural network based on rough set theory. At last, an example of off-line handwritten numeral identification attests the validity of this method.