Mutual Coupling Reduction in Antenna Arrays Using Artificial Intelligence Approach and Inverse Neural Network...

Area of Science:

• Electromagnetics and Wave Theory
• Antenna Array Design
• Metamaterials and Resonator Applications

Background:

• Mutual coupling in antenna arrays degrades overall system performance.
• Traditional methods for coupling reduction can be complex and time-consuming.
• Optimizing resonator parameters for coupling suppression requires efficient modeling techniques.

Purpose of the Study:

• To present a novel approach for reducing mutual coupling in antenna arrays.
• To introduce the use of custom-designed stepped-impedance resonators for isolation enhancement.
• To demonstrate the application of inverse artificial neural network (ANN) surrogate modeling for optimizing resonator geometry.

Main Methods:

• Design and fabrication of two patch antenna cells operating at 2.45 GHz.
• Integration of a stepped-impedance resonator between antennas to mitigate coupling.
• Development of an inverse ANN surrogate model trained with particle swarm optimization (PSO) using electromagnetic simulation data.
• Directly obtaining optimum resonator dimensions from target S-parameter values.

Main Results:

• Achieved isolation of -46.2 dB at 2.45 GHz, a significant improvement over conventional setups.
• Demonstrated remarkable cancellation of surface currents between antennas at the operating frequency.
• Validated the effectiveness of the inverse ANN surrogate model in optimizing resonator parameters.
• Showcased accelerated design process due to surrogate modeling, avoiding direct electromagnetic optimization.

Conclusions:

• The proposed method effectively reduces mutual coupling in antenna arrays using custom resonators and inverse surrogate modeling.
• Inverse ANN surrogate modeling offers an efficient and accelerated pathway to optimize resonator designs for enhanced antenna isolation.
• The achieved isolation levels represent a substantial advancement in antenna array performance.