Model-Based Mitigation of the Moving RFI in Radio Astronomy

Radio astronomical data are increasingly corrupted by human telecommunication activities. Therefore, Radio Frequency Interference (RFI) mitigation becomes an important step in the data processing flow. This paper considers the problem of adaptive array processing for interference canceling to drive nulls in a moving interference environment. As to the telescopes based on antenna array, RFI mitigation can be conducted in spatial domain using the sampled covariance matrix. In many practical scenarios, the achievable null depth is limited by covariance matrix estimation error which leads to poor identification of the interference subspace. We pay attention to the particularly troublesome cases of relatively rapid interference motion. A polynomial-based model is incorporated in the proposed algorithm to track changes in the array covariance matrix over time, mitigate interference subspace estimation errors, and improve canceler performance. Performance for conventional subspace projection (SP) is compared with polynomial-augmented SP using simulation. The simulation results validate the effectiveness of the proposed algorithm.