[Paper Review] A Simple AoA Estimation Scheme

Ahmed Badawy, Tamer Khattab+3 more|arXiv (Cornell University)|Sep 19, 2014

Indoor and Outdoor Localization Technologies17 references11 citations

TL;DR

This paper proposes a low-complexity, hardware-friendly angle-of-arrival (AoA) estimation scheme using a switched beam antenna array and cross-correlation with an omni-directional reference signal. The method achieves superior performance over the MUSIC algorithm, especially at low SNR levels (down to -25 dB), with resolution limited by beamwidth rather than SNR, and requires no prior knowledge of signal structure or number of sources.

ABSTRACT

We propose an intuitive, simple and hardware friendly, yet surprisingly novel and efficient, received signal's angle of arrival (AoA) estimation scheme. Our intuitive, two-phases cross-correlation based scheme relies on a switched beam antenna array, which is used to collect an omni-directional signal using few elements of the antenna array in the first phase. In the second phase, the scheme switches the main beam of the antenna array to scan the angular region of interest. The collected signal from each beam (direction or angle) is cross correlated with the omni-directional signal. The cross-correlation coefficient will be the highest at the correct AoA and relatively negligible elsewhere. The proposed scheme simplicity stems from its low computational complexity (only cross-correlation and comparison operations are required) and its independence of the transmitted signal structure (does not require information about the transmitted signal). The proposed scheme requires a receiver with switched beam antenna array, which can be attached to a single radio frequency chain through phase shifters, hence, its hardware friendliness. The high efficiency of our system can be observed by comparing its performance with the literature's best performing MUSIC algorithm. The comparison demonstrates that our scheme outperforms the MUSIC algorithm, specially at low SNR levels. Moreover, the number of sources that can be detected using our scheme is bound by the number of switched beams, rather than the number of antenna elements in the case of the MUSIC algorithm.

Motivation & Objective

• Address the high hardware and computational complexity of adaptive array systems (AAS) used in AoA estimation.
• Overcome the limitations of switched beam systems (SBS), such as poor performance at low SNR and inability to resolve closely spaced sources.
• Develop a system that maintains high performance at low SNR while reducing hardware complexity compared to conventional AAS methods.
• Eliminate the need for prior knowledge of the number of sources or signal correlation constraints, which are required by state-of-the-art methods like MUSIC.
• Provide a practical, hardware-friendly solution suitable for real-time deployment with a single RF chain and minimal baseband processing.

Proposed method

• Use a switched beam antenna array (SBA) to first collect an omni-directional signal as a reference, using a subset of elements (M₀) in the first phase.
• In the second phase, steer the main beam to scan the angular region of interest, collecting signals from multiple predefined beam directions.
• Cross-correlate each beam's received signal with the omni-directional reference signal to compute correlation coefficients.
• Identify the AoA as the direction corresponding to the maximum cross-correlation coefficient, which occurs only at the true angle of arrival.
• Leverage the fact that cross-correlation peaks at the correct AoA and is negligible elsewhere, enabling robust estimation without signal structure knowledge.
• Use Dolph-Chebyshev excitation for beamforming to achieve narrow main lobe beamwidth (HPBW), improving angular resolution independently of SNR.

Experimental results

Research questions

• RQ1 Can a simple cross-correlation-based scheme outperform the MUSIC algorithm in AoA estimation at low SNR levels?
• RQ2 Does the proposed scheme eliminate the need for prior knowledge of the number of sources or signal correlation properties?
• RQ3 To what extent does the resolution of the proposed system depend on SNR versus beamwidth, compared to MUSIC?
• RQ4 How does the number of collected samples affect the probability of false rejection (PFR) in the proposed scheme?
• RQ5 Can the system achieve high performance with a single RF chain and low computational complexity, making it suitable for hardware-constrained environments?

Key findings

• The proposed scheme achieves a probability of false rejection (PFR) of 5 dB at SNR = -25 dB, significantly outperforming the MUSIC algorithm, which fails at SNR levels below -20 dB with PFR = 3 dB.
• At SNR = 0 dB, the proposed scheme achieves improved performance with only 100 samples, and performance further improves with 1000 samples, demonstrating convergence with increased integration time.
• The resolution of the proposed system is consistently limited by the beamwidth (6° HPBW in the circular array case), unlike MUSIC, whose resolution degrades from 8° to 20° as SNR drops from 0 dB to -15 dB.
• The system can detect up to K sources, where K is the number of switched beams, which can exceed the number of antenna elements M, unlike MUSIC, which is limited to M sources.
• The method requires no prior knowledge of the number of sources or signal correlation, and is independent of the transmitted signal structure, making it more robust and broadly applicable.
• The computational complexity is minimal—limited to cross-correlation and comparison operations—making it significantly more efficient than MUSIC, which requires computationally intensive eigen-decomposition of the covariance matrix.