论文部分内容阅读
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 new modeling method based on artificial neural network is proposed, which provides a new approach for the modeling and detection of difficult parameters in complex systems. It is expected to replace the traditional mathematical modeling methods. According to the multi-layer forward network structure And improved BP algorithm, a black box model of nonlinear filter is established and the nonlinear triad stimulus nonlinear error of the color-sensitive detector is successfully solved. The simulation and experimental results show the feasibility of this method.