摘要
在测量系统中,传感器的非线性特性是测量系统误差的主要来源。要提高测量系统的精度,就必须进行误差补偿。设计了一个用模糊小脑神经网络实现的补偿环节。该补偿环节是一个用神经网络拟合的传感器逆特性,通过传感器的逆特性将传感器非线性特性改造成与实际物理过程相一致的不失真的线性特性,从而减小非线性误差。通过应用实验,验证了该方法的有效性。
The non-linear characteristic of transducer is the main resource of the error in measuring system. If the accuracy of the measuring system would be enhanced, the error had to be compensated. The compensation element was designed by the method of the fuzzy cerebellum neural network. The compensation element is an inverse characteristic of sensor fitted by neural network. Through the inverse characteristic of transducer, the non-linear characteristic of transducer is changed into linear characteristic without distortion, and identical to real physical process; thus non-linear error is decreased. The practice proves that the method is efficient.
出处
《自动化仪表》
CAS
2006年第3期11-13,17,共4页
Process Automation Instrumentation
关键词
非线性特性
误差
模糊
小脑神经网络
Non-linear characteristic Error Fuzzy Cerebellum neural network
