Improved Empirical Formula Modeling Method Using Neuro-Space Mapping for Coupled Microstrip Lines.

Area of Science:

• Electrical Engineering
• Computational Electromagnetics
• Microwave Engineering

Background:

• Coupled microstrip lines are fundamental components in high-frequency circuits.
• Accurate modeling is crucial for device performance and simulation efficiency.
• Existing empirical models often involve slow trial-and-error processes.

Purpose of the Study:

• To propose an improved empirical formula modeling method for coupled microstrip lines.
• To enhance model accuracy and reduce variable dependency using mapping neural networks (MNNs).
• To accelerate the optimization process through integrated sensitivity analysis.

Main Methods:

• Utilizing neuro-space mapping (Neuro-SM) for empirical formula modeling.
• Employing mapping neural networks (MNNs) with geometric and frequency variables.
• Incorporating simple sensitivity analysis expressions into the training process.

Main Results:

• The proposed Neuro-SM model accurately reflects the performance of coupled microstrip lines.
• The model achieves high accuracy with fewer variables compared to traditional methods.
• The integrated sensitivity analysis significantly accelerates the optimization process.

Conclusions:

• The developed Neuro-SM model provides an accurate and efficient method for coupled microstrip line analysis.
• The model demonstrates superior compatibility compared to existing simulation software.
• This approach offers a promising alternative for high-frequency circuit design and simulation.