Adaptive selective sidelobe canceller beamformer with applications in radio astronomy

Ronny Levanda, Amir Leshem

Abstract

We propose a new algorithm for parameter estimation that is applicable to imaging using moving and synthetic aperture arrays. The new method results in higher resolution and more accurate estimation than commonly used methods when strong interfering sources are present inside and outside the field of view, such as terrestrial interference or confusing astronomical sources.

Introduction

In radio astronomy, synthetic aperture imaging utilizes the rotation of the Earth to change the geometry of antenna arrays relative to celestial sources. This allows for high-resolution mapping of the sky. However, real-world antennas possess "sidelobes"—regions of sensitivity outside the main beam. When a powerful interference source, whether terrestrial or a bright "confusing" star, enters a sidelobe, its signal leaks into the data.

The standard approach to mitigate this is temporal averaging. However, because the array rotates, the impact of interference on a specific pixel is time-dependent. A pixel might be "clean" at one moment and heavily "polluted" minutes later as a sidelobe sweeps across an interferer. Averaging these epochs together effectively poisons high-quality data with noise, leading to artifacts and reduced resolution in the final "dirty image."

The Adaptive Selective Sidelobe Canceller (ASSC)

The Adaptive Selective Sidelobe Canceller (ASSC) shifts the paradigm from global averaging to local selection. Instead of treating all time epochs equally, ASSC identifies and combines only the "cleanest" observations for every specific pixel in the image.

Methodology

The ASSC algorithm operates in four primary stages:

  1. Epoch-wise Processing: The output power for every pixel $(l, m)$ is calculated for each time epoch separately using a beamformer, such as the classic Bartlett or the Minimum Variance Distortionless Response (MVDR) beamformer.
  2. Order Statistics Selection: For each pixel, the power values across all observed epochs are sorted.
  3. Selective Integration: The algorithm selects a subset of the smallest power values (the $\tilde{k}$ "best" epochs). The core heuristic is that while a true astronomical signal remains relatively constant, interference power fluctuates wildly as sidelobes sweep past. The minimum power values are the most likely to represent the true signal plus thermal noise, free from transient interference.
  4. Weighted Summation: These selected values are weighted and summed to produce the final ASSC image.
Beamformer power fluctuations across time epochs for sources $S_1$ and $S_2$.

As shown in Figure 4, the measured power at the locations of sources $S_1$ and $S_2$ varies significantly over time. Standard beamformers often yield values far exceeding the "True intensity" (the green dashed line) due to interference leakage. ASSC leverages the fact that the true intensity is found at the floor of these fluctuations.

Experimental Results

The algorithm was validated using simulated datasets featuring a 20-antenna array and a high-interference scenario where an out-of-field source was $10^6$ times (60 dB) stronger than the target astronomical sources. Simulations covered 12-hour observation periods with measurements taken every 1 to 6 minutes.

In these challenging conditions, standard Bartlett and MVDR beamformers failed, producing images dominated by sidelobe smearing. In contrast, the ASSC-enhanced MVDR beamformer successfully recovered the point sources. The researchers found that using as little as the best 5% of epochs was often sufficient to produce a superior image, significantly improving spatial resolution and contrast compared to traditional CLEAN-based approaches.

Conclusion and Future Work

ASSC provides a robust framework for enhancing existing beamforming techniques in radio astronomy. By intelligently selecting data based on temporal "cleanliness" at the pixel level, it effectively suppresses strong interference that would otherwise degrade image quality.

While the algorithm excels at cutting interference, picking only the minimum power values could potentially increase sensitivity to downward fluctuations in thermal noise if $\tilde{k}$ is too small. Future work will focus on automating the selection of the integration threshold ($\tilde{k}$) and optimizing computational efficiency for large-scale surveys, as sorting power values for every pixel across all epochs adds significant computational load.

Summary updated

Abstract

We propose a new algorithm, for parameter estimation that is applicable to imaging using moving and synthetic aperture arrays. The new method results in higher resolution and more accurate estimation than commonly used methods when strong interfering sources are present inside and outside the field of view (terrestrial interference, confusing sources).
BibTeX
@misc{levanda2010adaptive,
  author = {Ronny Levanda and Amir Leshem},
  title = {{Adaptive selective sidelobe canceller beamformer with applications in radio astronomy}},
  year = {2010},
  eprint = {1008.5066},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/1008.5066},
}
Figure