Ionogram RFI Rejection Using an Autoregressive Interpolation Process

Defence Science and Technology Group's Digital Oblique Receiving System (DORS) is a direct‐digital high‐frequency ionosonde receiver, capable of collecting high‐quality ionograms simultaneously on multiple oblique ionospheric paths. One of the key signal processing steps that run on board the receiver is a novel technique for the detection and removal of radio frequency interference (RFI). In the down‐converted narrow‐band time series, external RFI sources manifest as impulsive noise, often many orders of magnitude stronger than the ionosonde signal of interest. The RFI rejection technique applies a threshold detector to the whitened data, to locate the corrupt samples, and then replaces the bad data using forward and backward linear prediction, based on an autoregressive model of the desired ionosonde signal. Algorithm performance is evaluated using a set of simple statistics and ~6,000 DORS ionograms from central Australia. Compared to a more traditional approach of clipping the strong RFI impulses, the new technique significantly reduces instances of undersuppression and recovers more of the weaker propagation mode content in the ionogram. The result is a much cleaner image for the DORS automatic feature extraction and parameterized fitting technique. Plain Language Summary Ionospheric sounders (or ionosondes) are a widely used instrument for measuring the vertical electron density structure of the ionosphere, an outer charged layer of the Earth's atmosphere that refracts high‐frequency (HF) radio waves. An ionogram is a record of the path losses and delays for HF propagation via the ionosphere. However, ionograms can be potentially corrupted by the transmissions from other HF users, many of which are stronger than the ionosonde signal itself. This paper presents a novel signal processing technique for removing the influence of radio frequency interference in ionogram observations by Defence Science and Technology Group's new digital ionosonde. The result is a much cleaner standard of ionogram than traditionally seen, with improved sensitivity for detecting weaker propagation modes and fewer artifacts to degrade the automatic extraction of ionospheric parameters. Key Points An algorithm for detecting and rejecting radio frequency interference in ionosonde data has been developed and tested on a new ionosonde The algorithm detects impulses due to fixed‐frequency interferers and replaces these with an autoregressive model of the desired signal Performance