Neural network-based HMET scattering parameter extraction method and system, and storage medium

Abstract

The invention discloses a neural network-based HMET scattering parameter extraction method and system and a storage medium, and a storage medium. The extraction method comprises the steps: obtaining aplurality of groups of scattering parameter sample data of a high-electron-mobility transistor; extracting partial groups in the plurality of groups of scattering parameter sample data to train a neural network, wherein the neural network comprises a first neural network and a second neural network which are separated from each other; testing the remaining groups in the plurality of groups of scattering parameter sample data to obtain a neural network; when the neural network test is passed, obtaining test parameters of the to-be-tested high-electron-mobility transistor and inputting the testparameters into the neural network to output scattering parameters of the to-be-tested high-electron-mobility transistor; otherwise, continuing to train the neural network until the neural network passes the test.T he scattering parameters of the high-electron-mobility transistor are extracted by adopting the two independent neural networks, so that the speed is high, the accuracy is high, and the cost is relatively low. The method is widely applied to extraction of the scattering parameters of the high-electron-mobility transistor.

Description

technical field [0001] The invention relates to the extraction of scattering parameters of high electron mobility transistors, in particular to a neural network-based HMET scattering parameter extraction method, system and storage medium. Background technique [0002] High electron mobility transistor (High electron mobility transistor, HMET): A field effect transistor that utilizes the high mobility characteristics of a two-dimensional electron gas in a heterojunction or modulated doped structure. The HMET structure in this paper is a modulation-doped heterojunction, which forms an electronic potential well (approximately triangular) with the intrinsic semiconductor at its interface, and the electrons in the potential well are high-mobility two-dimensional electron gas (2-DEG ). Electrons are not scattered by ionized impurities in the potential well, so the electron mobility is high. Because HMET has the characteristic of two-dimensional electron gas, which makes it have ...

AI Technical Summary

Problems solved by technology

[0014] The above-mentioned current S-parameter extraction method uses a two-port network analyzer for testing. The process is cumbersome and complicated, and the test time is long. In addition, improper operations during the experimental test process may also lead to large errors in test results.

At the same time, the equipment of the network analyzer is expensive, and it needs to be used by professionals after training and learning, and the material cost and labor cost are too high

Images