On the Parameterization of Single Pole Adaptive Notch Filter Against Wide Range of Linear Chirp Interference

Abstract

Radio frequency interferences (RFIs) pose a severe threat to the Global Navigation Satellite System (GNSS). It is one of the main concerns that continue to challenge GNSS deployment in safety-related applications. Adaptive Notch Filter (ANF) is a classical method often used for the suppression of chirp signals. However, ANF performance is conditioned to the parameters selection where an inappropriate choice could also have a negative consequence. The adaptation step and pole contraction factor are mostly the two important parameters controlling the filter dynamics and suppression level. An appropriate compromise is needed while keeping in view the characteristics of the chirp signal. This study proposes a parameterization approach for ANF while dealing with a wide range of chirp signals with different values of sweep bandwidth, sweep rate, and power level. To provide labels RMSE criterion is investigated at the signal level before the de-spreading process. Finally, a generalized multivariate polynomial regression (MPR) model is presented that takes sweep bandwidth, sweep rate and power level as input features and predicts the values of the pole contraction factor and adaptation step. To validate the performance of ANF with the predictions, three scenarios with different chirp signals exhibiting slow, moderate and fast variations are presented. The mitigation performance is analyzed on several levels including, interference frequency estimation, satellite signal tracking, carrier-to-noise ratio and most importantly positioning KPIs such as accuracy, availability and safety.