论文部分内容阅读
包含新技术、新材料的非传统器件的不断涌现使现有的模型已不能完全表征THz器件的特性。而采用神经网络建模的方法,可极大地提高建模的效率和精确度,解决一系列传统模型所无法解决的问题,是一种新型的CAD建模方法。本文采用神经网络空间映射的方法,在传统的粗模型的基础上对输入信号进行有效地修正,从而得到适合太赫兹器件的精确模型,器件的截止频率Ft和最高振荡频率Fmax分别为220GHz和310GHz。模型在直流IV和1-110GHz范围内的S参数与测试结果吻合较好,比传统粗模型的精度有了较大的提高。
The continuous emergence of non-traditional devices that include new technologies and new materials has made the existing models unable to completely characterize THz devices. The use of neural network modeling method can greatly improve the efficiency and accuracy of modeling, to solve a series of traditional models can not solve the problem, is a new type of CAD modeling method. In this paper, the method of spatial mapping based on neural network is used to correct the input signal effectively on the basis of the traditional coarse model to get an accurate model for the terahertz device. The cut-off frequency Ft and the maximum oscillation frequency Fmax of the device are 220GHz and 310GHz . The S-parameters of the model in DC IV and 1-110GHz range are in good agreement with the test results, and the accuracy of the model is greatly improved compared with the traditional rough model.