Wiener Disorder Detection Method for Anti-Spoofing in GNSS Navigation Kalman Filters

This paper describes the adaptation of an algorithm, originally designed to detect a ramp deviation in a Wiener process, to spoofing detection in global navigation satellite system (GNSS, e.g., GPS) receivers. Imperfectly compensated spoofing signals exhibit common-mode errors across all signals broadcast by a single transmitter, driven by inaccurate compensation of the spoofer-to-receiver transmission channel (range and range rate). Those common-mode delay and frequency errors can affect the clock bias and frequency estimates that the victim receiver calculates to estimate its position, velocity, and time (PVT). In a benign situation (i.e., no spoofing), the change in clock bias and frequency drift over time is determined primarily by the receiver’s oscillator. Given that frequency drift is typically modeled as a Wiener process, our hypothesis was that the algorithm in [1] could be adapted to detect spoofer-induced deviations of clock frequency drift from the nominally expected stochastic behavior. First, the Wiener process ramp deviation algorithm is reviewed and then its application to spoofing detection in GNSS Kalman filters is discussed and formulated. Several practical issues are addressed, such as finite memory, computer processing limitations, and efficient approaches to reduce redundant calculations. Monte Carlo simulations characterizing false alarm and spoofing detection performance are presented. Finally, conclusions and recommendations for future work are presented.