Coarsely Quantized Spectral Estimation of Radio Astronomic Sources in Highly Corruptive Environments

摘要

This chapter is devoted to an application of non-Gaussian signal detection. In radio astronomy, given the huge flow of processed data, quantized correlators are widely used. Unfortunately, the occurrence of non-Gaussian interference can strongly alter the shape of the estimated spectra. In this chapter, the effect of a sine wave interference on a 1-bit correlator is first analysed. Then, a new interference detection criterion is proposed. It uses the real-time capabilities of quantized correlators and makes a statistical comparison between contaminated and non-contaminated quantized correlation functions. Thus, the final spectral estimation can be preserved by blanking the correlator in real-time. Simulations, using synthetic and actual data, are presented. This technique of real time detection can significantly improve the quality of spectral line observations.