摘要
针对光栅传感器输出细分方法的一些不足,从另一个角度研究了一种在不增加复杂的硬件电路和光栅刻线的情况下,采用软件实现对光栅传感器输出细分的方法。该方法将遗传算法与动态神经网络相结合,通过少量样本点的学习,用神经网络拟合出光栅传感器的传输特性。对任意输入值,经神经网络泛化后,实现对光栅传感器输出细分。通过实验验证了该方法的有效性。
Taking into account the flaw that output subdivision of grating transducer currently, this paper studies a new method that implement output subdivision with genetic algorithm and neural networks. The research shows that the networks can close with transducer input and output characteristic through learning by a very few sample data, when any data is imported, the networks can bring forth high accuracy subdivision output. The results of simulation show that the method is availability.
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2003年第9期1138-1140,共3页
Systems Engineering and Electronics
关键词
光栅传感器
细分
遗传算法
神经网络
Grating transducer
Subdivision
Genetic algorithm
Neural networks
