Gigahertz Bandwidth and Nanosecond Timescales: New Frontiers in Radio Astronomy Through Peak Performance Signal Processing
<p>Abstract In the past decade, there has been a revolution in radio-astronomy signal processing. High bandwidth receivers coupled with fast ADCs have enabled the collection of tremendous instantaneous bandwidth, but streaming computational resources are struggling to catch up and serve these...
| Summary: | <p>Abstract In the past decade, there has been a revolution in radio-astronomy signal processing. High bandwidth receivers coupled with fast ADCs have enabled the collection of tremendous instantaneous bandwidth, but streaming computational resources are struggling to catch up and serve these new capabilities. As a consequence, there is a need for novel signal processing algorithms capable of maximizing these resources. This thesis responds to the demand by presenting FPGA implementations of a Polyphase Filter Bank which are an order of magnitude more efficient than previous algorithms while exhibiting similar noise performance. These algorithms are showcased together alongside a broadband RF front-end in Starburst: a 5 GHz instantaneous bandwidth two-element interferometer, the first broadband digital sideband separating astronomical interferometer. Starburst technology has been applied to three instruments to date.</p>
<p>Abstract Wielding tremendous computational power and precisely calibrated hardware, low frequency radio telescope arrays have potential greatly exceeding their current applications. This thesis presents new modes for low frequency radio-telescopes, dramatically extending their original capabilities. A microsecond-scale time/frequency mode empowered the Owens Valley Long Wavelength Array to inspect not just the radio sky by enabling the testing of novel imaging techniques and detecting overhead beacon satellites, but also the terrestrial neighborhood, allowing for the characterization and mitigation of nearby sources of radio frequency interference (RFI). This characterization led to insights prompting a nanosecond-scale observing mode to be developed, opening new avenues in high energy astrophysics, specifically related to the radio frequency detection of ultra-high energy cosmic rays and neutrinos.</p>
<p>Abstract Measurement of the flux spectrum, composition, and origin of the highest energy cosmic ray events is a lofty goal in high energy astrophysics. One of the most powerful new windows has been the detection of associated extensive air showers at radio frequencies. However, all current ground-based systems must trigger off an expensive and insensitive external source such as particle detectors - making detection of the rare, high energy events uneconomical. Attempts to make a direct detection in radio-only data have been unsuccessful despite numerous efforts. The problem is even more severe in the case of radio detection of ultra-high energy neutrino events, which cannot rely on in-situ particle detectors as a triggering mechanism. This thesis combines the aforementioned nanosecond-scale observing mode with real-time, on-FPGA RFI mitigation and sophisticated offline post-processing. The resulting system has produced the first successful ground based detection of cosmic rays using only radio instruments. Design and measurements of cosmic ray detections are discussed, as well as recommendations for future cosmic ray experiments. The presented future designs allow for another order of magnitude improvement in both sensitivity and output data-rate, paving the way for the economical ground-based detection of the highest energy neutrinos.</p> |
|---|