CMAS Lab - Machine Learning Based Inverse Model for Identifying

Journal article

Mohd. Yusuf, Avirup Dasgupta, Sourajeet Roy
Electrical Design of Advanced Packaging and Systems Symposium, 2024

Cite

APA Click to copy
Yusuf, M., Dasgupta, A., & Roy, S. (2024). Machine Learning Based Inverse Model for Identifying Transmission Line Network Parameters from Tabulated Frequency-Domain Data. Electrical Design of Advanced Packaging and Systems Symposium.

Chicago/Turabian Click to copy
Yusuf, Mohd., Avirup Dasgupta, and Sourajeet Roy. “Machine Learning Based Inverse Model for Identifying Transmission Line Network Parameters from Tabulated Frequency-Domain Data.” Electrical Design of Advanced Packaging and Systems Symposium (2024).

MLA Click to copy
Yusuf, Mohd., et al. “Machine Learning Based Inverse Model for Identifying Transmission Line Network Parameters from Tabulated Frequency-Domain Data.” Electrical Design of Advanced Packaging and Systems Symposium, 2024.

BibTeX Click to copy

@article{mohd2024a,
  title = {Machine Learning Based Inverse Model for Identifying Transmission Line Network Parameters from Tabulated Frequency-Domain Data},
  year = {2024},
  journal = {Electrical Design of Advanced Packaging and Systems Symposium},
  author = {Yusuf, Mohd. and Dasgupta, Avirup and Roy, Sourajeet}
}

Abstract

In this paper, a neural network based inverse model is presented that can predict the values of the geometrical, physical, and material parameters of a given transmission line network from its tabulated frequency-domain responses. The proposed inverse model consists of two parts – a forward neural network that emulates the frequency-domain responses of the transmission line network under test as analytic functions of the geometrical, physical, and material parameters and a particle swarm optimizer (PSO) that minimizes an appropriate loss function to derive the requisite values of the network parameters.