Cyclostationary Detectors for RFI Mitigation in Radio Astronomy
Stephanie Bretteil (LESI/Polytech Orleans)
Rodolphe Weber (LESI/Polytech'Orleans)

This paper presents a scheme for removing efficiently cyclostationary radio frequency interferences (RFI) from astronomical data. Various methods have been experimented to eliminate those RFI depending on the type of interferences and the type of instruments. The present study focuses on time-frequency blanking on data coming from a single dish. Time-frequency blanking consists in removing polluted data blocks in the time-frequency plane before integration in order to clean up the spectrum. The time-frequency plane is obtained in real time by a digital filter bank based on FFT or polyphase filter. The blanking decision depends on the kind of detectors implemented.

The detection criteria are generally based on a statistical contrast between the signal of interest (SOI) and the RFI. For example, the most used criterion is the power estimation. Indeed, its low complexity makes it possible real time implementation. But, its effectiveness depends on the RFI level which must be obviously high. Moreover, the probability of false alarm may increase rapidly if the reference level is not stationary or if the SOI becomes so bright that the system may blank both the RFI and the SOI. This paper presents two detectors based on the temporal properties of a particular class of RFI, called the cyclostationary signals. These detectors are power independent. Besides they can work with real time constraints and detect RFI with a low Interference Signal Ratio (ISR). The first section will present the hypothesis and the properties of the signals used in this paper. The second one will define the cyclostationarity and will compare the two detectors. In the third section, an example of time-frequency blanking is described and discussed.