Scholarlink

This paper outlines the design and Radio Frequency (RF) performance comparison of both a traditional and an anti-biased RF Micro Electro Mechanical System (MEMS) Varactor. The anti-biased configuration is implemented for RF MEMS Varactor in a shunt arrangement. Through High Frequency Structure Simulator (HFSS) simulations, the RF performance of the anti-biased varactor is studied. The results demonstrate that the anti-biased varactor exhibits superior RF characteristics, whereas the position of the varactor determines isolation loss. Improvements in Isolation loss could be done using time consuming Electro Magnetic (EM) tools. Such a laborious process can be reduced using regressor models such as eXtreme Gradient boosting algorithms (XG-Boost). The objective of this study is to enhance RF performance by modifying physical parameters such as position (p) and frequency (f) using the XG-Boost technique, focusing on improving insertion loss (S21) and return loss (S11). Through the implementation of the XG-Boost method, varactor design time is drastically reduced by 72.73% compared to conventional EM simulators, which typically consume around 110 min on an Intel-Core (2.4 GHz) CPU equipped with 8 GB of RAM.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Optimization of RF MEMS Based Anti-biased Varactor Using Regressor Model

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • S. Harish,
  • R. Shreya,
  • R. Vishwadhika,
  • S. Kanthamani

摘要

This paper outlines the design and Radio Frequency (RF) performance comparison of both a traditional and an anti-biased RF Micro Electro Mechanical System (MEMS) Varactor. The anti-biased configuration is implemented for RF MEMS Varactor in a shunt arrangement. Through High Frequency Structure Simulator (HFSS) simulations, the RF performance of the anti-biased varactor is studied. The results demonstrate that the anti-biased varactor exhibits superior RF characteristics, whereas the position of the varactor determines isolation loss. Improvements in Isolation loss could be done using time consuming Electro Magnetic (EM) tools. Such a laborious process can be reduced using regressor models such as eXtreme Gradient boosting algorithms (XG-Boost). The objective of this study is to enhance RF performance by modifying physical parameters such as position (p) and frequency (f) using the XG-Boost technique, focusing on improving insertion loss (S21) and return loss (S11). Through the implementation of the XG-Boost method, varactor design time is drastically reduced by 72.73% compared to conventional EM simulators, which typically consume around 110 min on an Intel-Core (2.4 GHz) CPU equipped with 8 GB of RAM.