Radio frequency filter modeling method

Abstract

The invention discloses a radio frequency filter modeling method, which comprises the steps of obtaining modeling sample data, preprocessing the sample data, establishing a neural network model, training the neural network model, packaging the neural network model, generating a packaging model packet, constructing a neural network model of the filter by using a BP neural network algorithm, and performing heterogeneous model packaging on the neural network model of the filter for professional radio frequency simulation software and tools to form a filter packaging model which can be directly called. According to the method, input and output data are subjected to adaptive training through a BP neural network algorithm to learn filter behavior characteristics, the complex nonlinear mapping relation of the model can be accurately fitted, the internal structure and the working principle of the filter do not need to be known, and the established filter neural network model is packaged through a heterogeneous model, and can be directly used by professional radio frequency simulation software and tools for system simulation. Compared with a traditional modeling method, the method has the obvious advantages.

Description

technical field [0001] The invention relates to the technical field of radio frequency / microwave modeling and simulation, in particular to a radio frequency filter modeling method. Background technique [0002] As an important part of RF / microwave modeling and simulation, RF filter modeling plays a key role in accurately evaluating the electrical performance and signal transmission performance of RF and microwave systems. [0003] Traditional RF filter modeling methods mainly include: physical model method, equivalent circuit model method and numerical method. However, the calculation speed of the physical model and the equivalent circuit model is fast, but it is necessary to understand the internal structure and working principle of the filter. When there is a lack of in-depth understanding of various parasitic effects and coupling effects, it is difficult to obtain an accurate model; numerical methods It takes a huge amount of calculation to get a more accurate model. In...

AI Technical Summary

Problems solved by technology

However, the calculation speed of the physical model and the equivalent circuit model is fast, but it is necessary to understand the internal structure and working principle of the filter. When there is a lack of in-depth understanding of various parasitic effects and coupling effects, it is difficult to obtain an accurate model; numerical methods It takes a huge amount of calculation to get a more accurate model

In addition, there is no effective heterogeneous model packaging method for the RF filter model established based on various theoretical methods, which makes it difficult for professional RF simulation software and tools (such as ADS software) to apply the established model to the system. Link simulation increases the difficulty of model reuse

Images