A Two-Dimensional (2-D) Sensor Network Architecture with Artificial Intelligence Models for the Detection of Magnetic...

Area of Science:

• Geophysics
• Sensor Networks
• Machine Learning

Background:

• Traditional one-dimensional (1-D) magnetometer arrays have limitations in spatial characterization of magnetic anomalies.
• Accurate estimation of target motion parameters is crucial for effective surveillance and monitoring.

Purpose of the Study:

• To develop and evaluate a two-dimensional (2-D) network of magnetometers for enhanced magnetic anomaly detection.
• To improve spatial characterization of anomalies and enable reliable estimation of target motion parameters.
• To integrate deep learning for automated detection and preliminary target tracking.

Main Methods:

• Development of a 2-D network of sensor nodes, each with a custom electronic system and a biaxial fluxgate magnetometer operating in null mode.
• Deployment of a 5x5 node array over a 12x12 m^2 area for terrestrial tests with moving ferromagnetic cylinders.
• Application of deep learning models to process magnetometer data for information extraction.

Main Results:

• Demonstrated the feasibility of the 2-D magnetometer network configuration.
• Showcased enhanced spatial characterization of magnetic anomalies compared to 1-D systems.
• Validated the capability for automated detection and preliminary tracking of moving targets.

Conclusions:

• The 2-D magnetometer network offers significant improvements for magnetic anomaly detection.
• The system is suitable for integration into intelligent, real-time surveillance systems.
• Potential applications include security and underwater monitoring.