Sparse planar arrays for azimuth and elevation using experimental data.

Area of Science:

• Signal Processing
• Array Signal Processing
• Array Geometry

Background:

• Dense uniform arrays present limitations in sensor count and resolution for direction-finding.
• Sparse arrays offer improved performance with fewer sensors, demonstrated in 1D applications.
• Compressive sensing algorithms like sparse Bayesian learning (SBL) are effective for sparse vector estimation.

Purpose of the Study:

• To investigate the efficacy of nested and co-prime sparse arrays for 2D direction-of-arrival (DOA) estimation.
• To propose and evaluate a multi-frequency variant of sparse Bayesian learning (SBL).
• To compare SBL performance against conventional beamforming and Multiple Signal Classification (MUSIC) using experimental data.

Main Methods:

• Utilized nested and co-prime sparse arrays derived from a rectangular array.
• Applied sparse Bayesian learning (SBL), a compressive sensing algorithm, for DOA estimation.
• Developed and implemented a multi-frequency SBL variant to handle non-flat source frequency spectra.
• Conducted experimental validation in an anechoic chamber using real data.

Main Results:

• Sparse arrays, specifically nested and co-prime configurations, demonstrated superior performance in 2D DOA estimation.
• The proposed multi-frequency SBL variant effectively processed multi-snapshot and multi-frequency data.
• Experimental results validated the numerical findings, showing improved resolution and accuracy compared to traditional methods.

Conclusions:

• Nested and co-prime sparse arrays combined with SBL provide a robust solution for 2D DOA estimation.
• The multi-frequency SBL approach enhances estimation accuracy, particularly for sources with varying spectral content.
• SBL with sparse arrays presents a promising alternative to conventional DOA estimation techniques in practical scenarios.