摘要
提出了一种基于人工神经网络的新的建模方法，为复杂系统中难测参数的建模和检测提供了新的途径，并有望取代传统的数学建模方法；根据多层前向网络结构和改进的BP算法，建立了一个非线性滤波器的黑箱模型，并成功地解决了色敏探测器的光谱三刺激值非线性误差问题；模拟与实验结果表明了这种方法的可行性。
A medeling method based on artificial neural network is put fowird, which can preduce a new way for the modeling and measuring of the parameters hard to measure in complex systems and hope to replace the traditional mathmahca methed. The black - box model established based on the back - propagation and nonlinear filter can successfully overeome the spectral tristimulus value's noulinear errors of cofor sensor. Both simulation and experimental results proe its feasibility.
出处
《自动化与仪器仪表》
1998年第5期8-10,共3页
Automation & Instrumentation
关键词
色差计
非线性误差
人工神经网络
Neural network Tristimulus value Noalinear filter
