Evaluation of the kurtosis algorithm in detecting radio frequency interference from multiple sources

Abstract

-A few of the issues faced by the kurtosis detection algorithm on recent field campaigns is discussed here. The performance of the kurtosis algorithm in detecting multiple-source Radio Frequency Interference (RFI) is characterized. A new RFI statistical model is presented in the paper to take into account the behavior of RFI sources under a large foot-print. Results indicate the behavior of the kurtosis ratio under central-limit conditions due to large number of RFI sources.

Surveys and Analysis of RFI in Preparation for SMOS: Results from Airborne Campaigns and First Impressions from Satellite Data

“…However, it has also become clear that the kurtosis method has its weaknesses, and sometimes TB spikes much larger than what can be ascribed to natural radiation, i.e., obvious RFI, go undetected. This has been noticed by the authors of this paper and also reported in [23].…”

“…Another possible explanation is given in [23]. If many RFI sources are seen at the same time by the radiometer system, then there is a tendency toward normality due to the central-limit theorem irrespective of the distribution functions of the individual sources.…”

Demonstration of a Combined Active/Passive L-Band Microwave Remote Sensing Airborne Instrument

“…• the power spectrum is obtained using a polyphase filter bank (PFB) [33], while in the previous PALS version the spectrum information was not available, • RFI detection and mitigation techniques such as Kurtosis [34], [35], [36] are applied, while no mitigation strategy was implemented before, and • the substitution of the previous discrete components mature technology by integrated components reducing the size and weight of the present design by an order of magnitude, from 4 racks to a single one.…”