晶体振荡器老化预测的径向基函数神经网络法

A Neural Networks Method of Radial Basis Function for Aging Prediction of Quartz Crystal Oscillator

建立了预测石英晶体振荡器老化的一种径向基函数神经网络模型,这种人工神经网络由输入层和输出层组成,输入层由计算径向距离范数的非线性神经元组成,输出层由一个计算径向基函数的神经元组成.提出了确定规格化径向距离尺度因子的一种方法,并在此基础上导出了一种径向基函数神经网络的学习算法,这种算法具有计算形式简单和易于实现的优点,适合于用加速老化法和外推法进行石英晶体振荡器老化预测的实验数据处理.

A neural network model of radial basis function for aging prediction of quartz crystal oscillator is presented, it is composed by the input layer and the output layer. The input layer consists of non-linear neuron for calculating the radial distance norm, and the output layer consists of neuron for calculating the radial basis function. A determination of standardization radial distance scale factor is put forward, and in this foundation, a determination of the neural network learning algorithm of radial basis function is proposed. Also several examples of the results obtained by this method are given.