Machine Learning-Assisted Design and Optimization of On-Chip Inductor

  • Research field:RFIC Design, Passive Component Modeling
  • Type:Bachelor/Masterarbeit
  • Time:anytime
  • Supervisor:

    M.Eng Hapsah Aulia Azzahra

  • Note:

    in English

  • Integrated inductor plays an important role in impedance matching, frequency tuning, etc., in many RFIC building blocks. An inductor performance is usually defined by many geometrical parameters. Traditional inductor design involves time-consuming trial-and-error using layout and electromagnetic (EM) simulation tools. This study aims to develop a machine learning-assisted tool for automatic synthesis of on-chip inductor, targeting RF applications. Unlike the traditional design flows, this approach uses modern machine learning to achieve the goals of inductor design.

    Task
    - Build a dataset of inductor geometries.
    - Train and optimize machine learning models to predict inductor performance from geometry.
    - Validate synthesized design using EM simulations.

    Requirements
    - Good understanding of passive devices.
    - Machine learning knowledge is preferred.