OSKAR: Simulating digital beamforming for the SKA aperture array

Mort, Ben; Dulwich, Fred; Salvini, Stefano; Adami, Kristian Zarb; Jones, M. E.

doi:10.1109/array.2010.5613289

Cited by 24 publications

(11 citation statements)

References 2 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To test the RTS method of sky generation, we therefore needed a fast and discretised method. Another excellent CUDA accelerated simulation package, OSKAR (Mort et al, 2010), addresses these two points. However, the RTS also generates parts of the sky model via shapelets (see Line et al, 2020 for an overview), which OSKAR cannot.…”

Section: Statement Of Needmentioning

confidence: 99%

WODEN: A CUDA-enabled package to simulate low-frequency radio interferometric data

Line¹

2022

JOSS

View full text Add to dashboard Cite

WODEN is designed to simulate the response of a class of telescope known as an interferometer, producing output "visibilities" for a given astrophysical sky model. Simulated observations allow us to test other software packages that are designed to calibrate and analyse real interferometric data, including verifying expected behaviour with known inputs, and testing new sky modelling techniques. The WODEN sky model can be specified in Dirac-delta like functions on the sky (known in the field as "point sources"), elliptical Gaussian models, or built out of "shapelet" basis functions, allowing complicated morphologies to be created. Users are able to input a bespoke layout for the interferometer, vary a number of observational parameters including time of day, length of observation and frequency coverage, and select from a number of predefined primary beams which encode the response of the receiving elements of an interferometer. This allows simulations of a number of telescopes to be undertaken. WODEN works with input Stokes I, Q, U, V polarisations as a sky model, simulating telescopes with dual linear polarisations, and outputting linear Stokes polarisations.

show abstract

Section: Statement Of Needmentioning

confidence: 99%

WODEN: A CUDA-enabled package to simulate low-frequency radio interferometric data

Line¹

2022

JOSS

View full text Add to dashboard Cite

show abstract

“…The 512 stations are divided into two parts: (1) 224 stations are randomly distributed within the 'core' region of 1000 m in diameter; (2) the remaining stations are placed on three spiral arms extending up to a radius of about 35 km. For each sky map, we employ the OSKAR 4 simulator (Mort et al 2010) to perform 6-hour synthesis imaging to obtain the visibility data, from which the 'observed' image is created by the WSClean 5 imager (Offringa et al 2014). In order to emphasize the faint and relatively diffuse EoR signal, the natural weighting and baselines of 30-1000 wavelengths are utilized in the imaging process.…”

Section: Simulation Of the Ska Imagesmentioning

confidence: 99%

Separating the EoR signal with a convolutional denoising autoencoder: a deep-learning-based method

et al. 2019

Monthly Notices of the Royal Astronomical Society

View full text Add to dashboard Cite

When applying the foreground removal methods to uncover the faint cosmological signal from the epoch of reionization (EoR), the foreground spectra are assumed to be smooth. However, this assumption can be seriously violated in practice since the unresolved or mis-subtracted foreground sources, which are further complicated by the frequency-dependent beam effects of interferometers, will generate significant fluctuations along the frequency dimension. To address this issue, we propose a novel deep-learning-based method that uses a nine-layer convolutional denoising autoencoder (CDAE) to separate the EoR signal. After being trained on the SKA images simulated with realistic beam effects, the CDAE achieves excellent performance as the mean correlation coefficient (ρ) between the reconstructed and input EoR signals reaches 0.929 ± 0.045. In comparison, the two representative traditional methods, namely the polynomial fitting method and the continuous wavelet transform method, both have difficulties in modelling and removing the foreground emission complicated with the beam effects, yielding onlyρ poly = 0.296 ± 0.121 andρ cwt = 0.198 ± 0.160, respectively. We conclude that, by hierarchically learning sophisticated features through multiple convolutional layers, the CDAE is a powerful tool that can be used to overcome the complicated beam effects and accurately separate the EoR signal. Our results also exhibit the great potential of deep-learning-based methods in future EoR experiments.

show abstract

“…The drawbacks to testing on real data are numerous; however, a real instrument can introduce random and systematic errors; the ionosphere is a constantly changing source of error; the true sky brightness distribution is actually unknown, and consists of point like and diffuse emission with varying spectral behaviours. The latter is the biggest obstacle for this experiment, so simulated visibilities were created using OSKAR 11 (Mort et al 2010). In this manner, the input data could be completely understood and so the effects of positional inaccuracies be isolated.…”

Section: Simulations Of Foreground Removal Towards An Eor Detectionmentioning

confidence: 99%

PUMA: The Positional Update and Matching Algorithm

Line

Webster

Pindor

et al. 2017

Publ. Astron. Soc. Aust.

View full text Add to dashboard Cite

We present new software to cross-match low-frequency radio catalogues: the Positional Update and Matching Algorithm (PUMA). PUMA combines a positional Bayesian probabilistic approach with spectral matching criteria, allowing for confusing sources in the matching process. We go on to create a radio sky model using PUMA based on the Murchison Widefield Array Commissioning Survey, and are able to automatically cross-match 98.5% of sources. Using the characteristics of this sky model, we create simple simulated mock catalogues on which to test PUMA, and find that PUMA can reliably find the correct spectral indices of sources, along with being able to recover ionospheric offsets. Finally, we use this sky model to calibrate and remove foreground sources from simulated interferometric data, generated using OSKAR (the Oxford University visibility generator). We demonstrate that there is a substantial improvement in foreground source removal when using higher frequency and higher resolution source positions, even when correcting positions by an average of 0.3 given a synthesized beam-width of 2.3.Comment: Accepted for publication in Publications of the Astronomical Society of Australia (PASA) - 34 pages - 20 figures - PUMA github repository: https://github.com/JLBLine/PUM

show abstract

OSKAR: Simulating digital beamforming for the SKA aperture array

Abstract: Digital beamforming for the aperture array components of the SKA poses considerable computational challenges. In this paper, we propose a hierarchical scheme aimed at tackling them and introduce OSKAR, a beamforming simulator which implements these ideas and algorithms.

Cited by 24 publications

References 2 publications

WODEN: A CUDA-enabled package to simulate low-frequency radio interferometric data

WODEN: A CUDA-enabled package to simulate low-frequency radio interferometric data

Separating the EoR signal with a convolutional denoising autoencoder: a deep-learning-based method

PUMA: The Positional Update and Matching Algorithm

Contact Info

Product

Resources

About