The Magellanic Clouds have long provided a rich celestial laboratory for many astrophysical research programmes. Their location relatively close to the Earth and away from the plane of our Galaxy has made them a natural target for Southern Hemisphere ground-based instrumentation. Likewise, the continuing quest for images of the Clouds with higher dynamic range and improved angular resolution has driven a continual improvement in instrumentation across a range of wavelength bands. The cornerstone of this thesis is a study of selected sources in the Magellanic Clouds. The sample was chosen from the 843MHz Molonglo Observatory Synthesis Telescope survey of the Clouds, based on the existing knowledge of each source, its flux density and angular extent. This sample was used to explore observational and analysis techniques with the Australia Telescope Compact Array in order to better determine the nature of these objects and to identify those sources worthy of further study. This work highlights many pertinent issues associated with the correct classification of sources when only a limited amount of data is available. These issues led directly to the development of a more systematic approach in the classification of the Large Magellanic Cloud source sample, detailed for the first time in this thesis. Two supernova remnants in the Small Magellanic Cloud were studied in detail. The Australia Telescope images of 1E0102.2-7219 revealed, for the first time, the radio structure of this young oxygen-rich supernova remnant, and allowed a detailed comparison with existing optical and X-ray data to be undertaken. The comparisons presented in this thesis and in an earlier publication have prompted exciting new X-ray observations at unprecedented angular resolution. The second, 0101-7226, studied as part of an international collaboration, has a shell morphology at radio wavelengths but no associated X-ray emission and is therefore something of an enigma.

The Magellanic Clouds have long provided a rich celestial laboratory for many astrophysical research programmes. Their location relatively close to the Earth and away from the plane of our Galaxy has made them a natural target for Southern Hemisphere ground-based instrumentation. Likewise, the continuing quest for images of the Clouds with higher dynamic range and improved angular resolution has driven a continual improvement in instrumentation across a range of wavelength bands. The cornerstone of this thesis is a study of selected sources in the Magellanic Clouds. The sample was chosen from the 843MHz Molonglo Observatory Synthesis Telescope survey of the Clouds, based on the existing knowledge of each source, its flux density and angular extent. This sample was used to explore observational and analysis techniques with the Australia Telescope Compact Array in order to better determine the nature of these objects and to identify those sources worthy of further study. This work highlights many pertinent issues associated with the correct classification of sources when only a limited amount of data is available. These issues led directly to the development of a more systematic approach in the classification of the Large Magellanic Cloud source sample, detailed for the first time in this thesis. Two supernova remnants in the Small Magellanic Cloud were studied in detail. The Australia Telescope images of 1E0102.2-7219 revealed, for the first time, the radio structure of this young oxygen-rich supernova remnant, and allowed a detailed comparison with existing optical and X-ray data to be undertaken. The comparisons presented in this thesis and in an earlier publication have prompted exciting new X-ray observations at unprecedented angular resolution. The second, 0101-7226, studied as part of an international collaboration, has a shell morphology at radio wavelengths but no associated X-ray emission and is therefore something of an enigma.

Radio astronomical synthesis imaging arrays can create images with resolution much higher than can single dish telescopes. However, one of the biggest problems that imaging arrays face is radio frequency interference (RFI). This interference corrupts signals and prevents accurate image creation. Therefore, it is necessary to remove this interference. This thesis discusses the synthesis imaging procedure and array spatial filter­ing methods to remove interference, including Multiple Sidelobe Canceller (MSC), Subspace Projection (SP), and Cross­-Subspace Projection (CSP). The CLEAN algorithm, an image restoration technique, is also discussed. Various improvements to the VSA are discussed, including upgrades to the hardware and software and addition of a fifth antenna to the array. Calibration techniques for the VSA are presented. Successful image synthesis for deep-­space sources of Cassiopeia A and Cygnus A are shown and phase errors that have caused difficulties with imaging are considered. The previously mentioned algorithms are successfully applied to data gathered by the Very Small Array (VSA), allowing images to be created in environments with interference. An improved method for bias correction for both SP and CSP is demonstrated. The CLEAN algorithm is demonstrated on two different images.

Radio astronomy synthesis imaging arrays are composed of many parabolic reflector antennas. These antennas are designed to be extremely sensitive to detect faint emissions from astronomical sources. Unfortunately, this also makes them susceptible to radio frequency interference (RFI) from man made sources such as orbiting satellites. The radio astronomy research group at Brigham Young University (BYU) is investigating methods to mitigate the effects of RFI in radio astronomy synthesis imaging. Though real-time RFI mitigation has been demonstrated for a large single dish telescope, for synthesis imaging arrays our prior work has consisted solely of algorithm development and computer simulations. To test our algorithms on experimental data we need an image synthesis array at BYU. The primary contribution of this Master's thesis is the design and construction of a working image synthesis array on the roof of the Clyde Building at BYU. This thesis describes the design of the antenna placement for the synthesis array. Antenna placement is the primary factor for determining image quality since the placement dictates the shape of the synthesized beam. Simulations were performed, prior to the array's construction, to predict the quality of images from the array. Another contribution of this thesis is signal processing code to generate correlations of the signals from the antennas. Code was written to calibrate measured data and generate an image from the correlations. Code was also written to steer the antennas and track astronomical phenomena. The performance of the array is evaluated in this thesis. The culmination of this thesis is a radio image of the supernova remnant Cassiopeia A. This thesis concludes with simulations of an RFI mitigation experiment that can be performed with the new array (pending certain improvements to the array).

Radio Astronomy, the study of distant objects in the radio spectrum, is set at defiance against all other users of the electromagnetic spectrum. Where traditional use would fill the electromagnetic spectrum with as much transmission and information as possible, Radio Astronomy would rather no man-made signals existed. Since that is not possible, they grumble and have to deal with unwanted transmissions impinging upon their instruments. I have demonstrated that subspace projection can remove these unwanted signals from Radio Astronomical data in post processing. I demonstrate it both on simulated data and on data taken from the Very Large Array radio telescope.In the process of implementing the algorithm, I show that the self power terms contain necessary information about the array element responses to RFI sources. While the autocorrelation are not used in the normal synthesis imaging process, my work shows that RFI mitigation using Subspace Projection performs better with the autocorrelations retained when computing projection matrices. Furthermore, I demonstrate that proper data collection allows a significant decrease in error under subspace projection. Potential enhancements to subspace projection are also briefly considered.

The Brigham Young University (BYU) Radio Astronomy group, in collaboration with Cornell University, the University of Massachusetts, and the National Radio Astronomy Observatory (NRAO), have in recent years developed and deployed PAF systems that demonstrated the advantages of PAFs for astronomy. However, these systems lacked the necessary bandwidth and acquisition times to be scientifically viable. This thesis outlines the development of a 20-MHz bandwidth system that can acquire for much longer periods of time and across much larger bandwidths than previous BYU systems. A report of the deployment of this system on the 305-meter reflector at the Arecibo Observatory in Puerto Rico is also summarized.The Commonwealth Scientific and Industrial Research Organisation (CSIRO) is currently constructing a PAF-equipped synthesis imaging array named the Australian Square Kilometre Array Pathfinder (ASKAP) that offers great promise for widening FOVs and enhancing RFI mitigation techniques. Previous work in RFI mitigation has demonstrated effective cancellation for synthesis imaging arrays under the assumption that the processing bandwidth is narrowband and correlator dump times are short. However, these assumptions do not necessarily reflect real-world instrument limitations. This thesis explores simulated adaptive array cancellation algorithm effectiveness as applied on the ASKAP instrument given realistic bandwidths and correlator dump times. The results demonstrate that active RFI mitigation performed across long baselines is largely ineffectual.