摘要
提出了一种采用神经网络逆运算补偿传感器非线性误差的方法.该方法先通过静态标定得到实验数据,然后采用单输入/单输出的模糊小脑神经网络(SISO FCMAC)建立传感器静态非线性模型,再由SISO FC-MAC的逆运算建立静态逆模型.与直接用神经网络建立逆模型的补偿方法相比较,具有学习简单、精度高和可在线标定等优点,且算法可以在单片机上实现.最后,通过实验验证了该方法的有效性.
This paper presents a method to compensate non-linearity of sensor using. With static calibration of experimental data, the nonlinear model of the sensor is built on a Single-Input-Single-Output Fuzzy Cerebllar Model Articulation Controller (SIS0 FCMAC ) , and the inverse model is established by inverse operation of SIS0 FCMAC. This method is compared with the direct use of neural network inverse model compensation, with the advantages of simple study, high precision and of online calibration, and algorithm can be implemented on the MCU. Finally, the experimental results show that the scheme is effective.
出处
《南京师大学报(自然科学版)》
CAS
CSCD
北大核心
2010年第4期172-176,共5页
Journal of Nanjing Normal University(Natural Science Edition)
基金
国家自然科学基金(60774060)
关键词
测量
非线性特性
脑神经网络
逆运算
measurement, nonlinear characteristics, neural network, inverse operation