摘要
针对电池分选中,由于输入技术参数存在噪声而引起误选率过高的问题,以 BP(badk-propagation)神经网络为基础,使用抽样学习的方法,提出了多参数输入的电池分选方案.实验结果表明,该方法能在技术参数存在噪声的情况下,通过网络训练可减小分选误差.
In the course of battery-sorting, the ratio of wrong to correct is high because there is noise in input-parameter. In this paper, a new method for sorting batteries, in which sampling study is used, based on BP neural newtwork, is presented. The result explains that although there is noise in input-parameter, the ratio of wrong is decreased through sampling study, in the trained neural network.
出处
《哈尔滨理工大学学报》
CAS
2001年第5期10-13,共4页
Journal of Harbin University of Science and Technology
基金
国家"863"计划(715-004-0080)