GNSS Spoofer Localization via Moving Dual-Antenna Geometry: MCMC Particle Filtering Followed by Newton Refinement

GNSS spoofing poses serious threats to positioning, navigation, and timing (PNT) services. While spoofer localization is critical for threat elimination, existing methods often rely on costly antenna arrays or distributed receivers. This paper proposes a cost-effective localization method using a moving dual-antenna GNSS receiver, typically employed for attitude determination. The core idea is to utilize Single-Difference Carrier Phase (SDCP) measurements to impose geometric constraints on the spoofing signal’s direction. At each measurement epoch, these measurements define a set of conical surfaces, each corresponding to a possible integer ambiguity. By aggregating such constraints across multiple epochs, only the conical surfaces associated with the correct ambiguities intersect at a unique point, thereby enabling both ambiguity resolution and spoofer localization. To solve the mixed-integer nonlinear estimation problem, we adopt a Markov Chain Monte Carlo (MCMC) Particle Filter (PF) to infer spoofer locations and integer ambiguities jointly. A Newton-based refinement step is further integrated to improve estimation accuracy. Theoretical analysis quantifies the effects of various error sources on localization performance. Both simulation studies and real-world experiments validate the effectiveness and robustness of the proposed method.