GNSS Interference Detection and Geolocation from LEO Satellites – Satellite Formation and Payload Design Specific Considerations and Their Impact on the Detection Sensitivity and Geolocation Accuracy

The paper introduces the general concept of monitoring of GNSS interference from space. It highlights all the relevant functional building blocks which consists of the whole processing chain on the satellite from the earth-oriented antenna towards the analogue-to-digital Converter (ADC) and storage unit of the captured interference signals. The practical design considerations are mentioned. This consists of a preliminary design for a satellite constellation, including space segment and ground segment, which would be capable to demonstrate GNSS interference monitoring from space. Different constraints in terms of processing capabilities, available hardware resources for the detection and geolocation algorithm, satellite internal raw sample storage capacities, data link rates and number of ground stations are investigated, and a set of useful parameters are highlighted. The detection performance of algorithms across different domains are evaluated under different types of interferences namely continuous wave (CW), linear and non-linear chirp, and multiple interferences. Single satellite-based detection using Periodogram, Short-Time-Fourier-Transform (STFT), Smoothed-Pseudo-Wigner-Ville-Distribution (SPWVD) and dual satellite based cross-correlation detection methods are considered. Furthermore, the impact of the distance of a two satellite LEO formation on the geolocation accuracy is discussed. At first, the so-called narrow-band cross-ambiguity-function (CAF) for time-differential-of-arrival (TDoA) and frequency-differential-of-arrival (FDoA) estimation is explained and its limitations due of the fast changing behaviour of the line-of-sight (LOS) channels between the emitter and the LEO satellites. Within a simulation, the shape of the CAF is evaluated in both TDoA and FDoA domains for a specific kind of chirp interference signal. Afterwards, the TDoA and FDoA estimates are fed into an iterative weighed-least-squares (WLS) geolocation algorithm in order to determine the location of the emitter source. This results are compared with the estimator co-variance matrix of the geolocation algorithm. The estimator co-variance matrix consists of the design matrix and the co-variance matrix of measurement errors which are derived from the cramer-rao-lower-bounds (CRLB)s of TDoA and FDoA measurements. A statistical evaluation of the achievable geolocation accuracy for different emitter hypothesis within the common field-of-view (FOV) of the satellites is conducted and the most promising satellite distance in terms of achieving the best overall geolocation accuracy is shown.