ANN-Assisted Beampattern Optimization of Semi-Coprime Array for Beam-Steering Applications

Area of Science:

• Signal Processing
• Array Signal Processing
• Artificial Intelligence

Background:

• Semi-Coprime array with staggered steering (SCASS) offers adjustable parameters for beamforming.
• SCASS allows trade-offs between half-power beamwidth (HPBW) and peak-to-side-lobe ratio (PSLR).
• Optimizing these parameters for specific performance goals is challenging.

Purpose of the Study:

• To propose an artificial neural network (ANN) for estimating optimal SCASS parameters.
• To minimize HPBW for a given PSLR and steering angle.
• To enhance beamforming performance in SCASS systems.

Main Methods:

• Developed a dataset using Matlab simulations of SCASS beampatterns.
• Trained an ANN using array architectural parameters, steering angle, and desired PSLR as inputs.
• Validated the ANN's ability to predict optimal adjustable parameters.

Main Results:

• The trained ANN accurately suggests optimal SCASS parameters.
• Achieved minimum HPBW for a desired PSLR and steering angle.
• Demonstrated a mean absolute error of only 0.83% in parameter estimation.

Conclusions:

• The proposed ANN effectively optimizes SCASS array parameters.
• This approach balances HPBW and PSLR for improved beamforming.
• The ANN provides a precise and efficient method for SCASS performance tuning.