Skip to Main content Skip to Navigation

Robust Statistics Applied to Radio Astronomy : Radio Frequency Interference Mitigation and Automated Spectral Line Detection for Broadband Surveys

Abstract : The advent of a new generation of radio telescopes coupled with digital processing hardware have provided tremendous new opportunities for extensive studies of the Galactic and extra galactic environment. In parallel, the nowadays huge amount of data produced has generated specific challenges. Among others: to secure high-speed data capture, to temporarily store huge quantities of raw data, to upgrade processing to this new data profile, and to provide researchers practical tools of analysis.In this thesis I have focused on two issues. Firstly the mitigation of radio frequency interference (RFI), and secondly practical methods for automated blind detection of spectral lines. The algorithms I present have been developed using observational data from the decimetric 100m-class single-dish Nanc{c}ay Radio Telescope (France) acquired with the WIBAR broadband and high sampling rate spectrometer. Contrary to classical spectral data analysis, The algorithms I have developed take into account time-series of flux-density. In order to excise RFI and detect spectral lines, I chose to apply robust statistics to these time-series, i.e., estimators of location and scale as well as regression which are immune to statistical outliers.In the first part, I lay down the basics of robust statistics applied to radio astronomy, used for this project: after having recalled basics of classical statistics in the context of radio astronomy, I discuss the pros and cons of several robust estimators of location and scale, followed by an evaluation of options for robust regression pertaining to the aims of this project. In the second part, I first describe the technical properties of the WIBAR broadband receiver, and then the fundamental assumptions and choices I have made to setup the architecture of the RObust Elusive Line detection (ROBEL) post-processing software I have developed for this matter. In the third part, I present results first from observations of different sources with WIBAR, and second related to RFI mitigation. After having discussed current issues, I conclude with possible future developments in interferometry in the SKA context.
Complete list of metadatas

Cited literature [102 references]  Display  Hide  Download
Contributor : Abes Star :  Contact
Submitted on : Friday, June 26, 2020 - 1:05:11 PM
Last modification on : Saturday, September 26, 2020 - 4:54:11 AM


Version validated by the jury (STAR)


  • HAL Id : tel-02882095, version 1



Christophe Belleval. Robust Statistics Applied to Radio Astronomy : Radio Frequency Interference Mitigation and Automated Spectral Line Detection for Broadband Surveys. Astrophysics [astro-ph]. Université Paris sciences et lettres, 2019. English. ⟨NNT : 2019PSLEO009⟩. ⟨tel-02882095⟩



Record views


Files downloads