Neuro-genetic system for optimization of GMI samples sensitivity

Area of Science:

• Materials Science
• Electrical Engineering
• Sensor Technology

Background:

• Giant Magnetoimpedance (GMI) sensors offer high potential for ultra-weak magnetic field measurements.
• Optimizing GMI sensor sensitivity is crucial but traditionally relies on time-consuming empirical methods.
• Existing models inadequately capture the complex dependencies of GMI sensitivity on various parameters.

Purpose of the Study:

• To develop an automated system for maximizing the impedance phase sensitivity of GMI sensor elements.
• To address the limitations of empirical parameter optimization in GMI sensor development.

Main Methods:

• A novel neuro-genetic system combining a Multi-Layer Perceptron (MLP) Neural Network and a Genetic Algorithm was employed.
• The MLP neural network models the impedance phase response of GMI samples.
• The Genetic Algorithm utilizes the MLP model to efficiently search for parameters that maximize phase sensitivity.

Main Results:

• The neuro-genetic system successfully modeled the impedance phase sensitivity of GMI samples.
• The system automatically identified the optimal set of conditioning parameters (sample length, magnetic field, DC level, excitation frequency) for maximizing sensitivity.
• Validation using a dataset of four different GMI sample lengths confirmed the system's effectiveness.

Conclusions:

• The proposed neuro-genetic system provides an efficient and automated approach to optimize GMI sensor parameters.
• This method significantly reduces the time and effort required for GMI sensor development.
• The findings pave the way for enhanced sensitivity in GMI-based magnetic field sensing applications.