Radio Frequency Interference Detection and Mitigation for GNSS and Radio Astronomy Applications

In this work we investigate techniques to detect and mitigate radio frequency interference (RFI) in global navigation satellite systems (GNSS) and radio astronomy applications. In the context of GNSS, a nonnegative matrix factorisation (NMF)-based detector is proposed which provides competitive detection capability and low false alarm rates for narrow and wideband RFI signals. Further, the proposed method does not require fine-tuning of parameters to detect specific RFI types, which are unknown a priori, while maintaining low computational complexity. NMF-based frameworks for RFI mitigation named supervised and semi-blind NMF are also proposed, with the former relying on prior knowledge about the interference, and the latter not requiring prior information. The proposed schemes are able to suppress typical RFI types found in real-world environments, outperforming techniques commonly used in the literature in strong interference scenarios. Further, the proposed technique is able to suppress multiple, different types of RFI without any prior information about the interference. In the context of radio astronomy applications, we propose a time-frequency based approach to detect automatic dependent surveillance-broadcast (ADS-B) RFI signals in the received radio-astronomy signals. The proposed method employs a frequency template vector for ADS-B signal detection that achieves high detection rates while maintaining low false alarms. The proposed approach achieves superior performance using real-life signals when compared with methods widely used in this domain. Moreover, three NMF-based schemes for RFI suppression for radio astronomy applications are proposed, viz.: supervised NMF with frequency selection, supervised NMF with RFI template, and semi-blind NMF. The former makes use of ADS-B signal samples while its RFI template variant employs a frequency template instead, with the semi-blind framework precluding any prior information about the interference. Using signals captured by the Parkes radio telescope in Australia, we demonstrate the efficacy of the proposed approaches in detecting and mitigating RFI, and their impact on pulsar observations. Results show that promising RFI suppression performance can be achieved with minimal degradation on the pulsar observations, signalling that NMF can be a potential tool in the radio astronomy field.