论文部分内容阅读
根据动态校准实验结果建立传感器的动态数学模型,以研究传感器的动态性能,是动态测试的一个重要内容.讨论了递归神经网络模型在传感器动态建模中的应用,给出了递归神经网络模型的结构及相应的训练算法.由于其反馈特征,使得递归神经网络模型能获取系统的动态响应.该方法特别适用于传感器非线性动态建模,而且避免了传感器模型阶次的选择的困难.试验结果表明,应用递归神经网络对传感器进行动态建模是一种行之有效的方法.
According to the result of dynamic calibration experiment, the dynamic mathematical model of sensor is established to study the dynamic performance of the sensor, which is an important content of the dynamic test. The application of the recurrent neural network model in the dynamic model of the sensor is discussed, and the recursive neural network model Structure and corresponding training algorithm.The recursive neural network model can obtain the dynamic response of the system because of its feedback characteristics.This method is especially suitable for the nonlinear dynamic modeling of the sensor and avoids the difficulty of selecting the order of the sensor model.The experimental results It shows that the use of recursive neural network to dynamically model the sensor is an effective method.