An availability-aware multi-domain measurement filter for antenna-array GNSS/INS under concurrent multipath and distributed spoofing
Abstract
Multipath and spoofing are difficult to separate when both appear as biased range measurements. We develop a Multi-Domain Resilient Navigation Filter (MDRNF) for an antenna-array Global Navigation Satellite System/Inertial Navigation System deeply integrated receiver. For each satellite, the filter uses navigation innovations, inter-antenna carrier-phase residuals and tracking-loop features. The array model predicts the direct path, so a reflected component is allowed to disturb the spatial residual. Each available domain is described by a normalized scale-chi-square working density; the temporal–spectral coupling term is used only as a generalized-Bayesian regularizer. Missing spatial evidence is left out of the update instead of being filled with a nominal value. The resulting hypothesis weights adjust pseudorange and Doppler covariances, isolate a channel when spoofing evidence is strong, and activate a conditional pseudorange-bias tracker. The receiver is evaluated in a 300 s MATLAB software-in-the-loop scenario with five antenna elements and eight GPS L1 C/A satellites. The scenario covers nominal tracking, satellite-dependent multipath, three distributed spoofers and recovery. An 80-run intermediate-frequency (IF) signal-level Monte Carlo simulation is used, with the IF signal and downstream receiver processing regenerated independently for every run. During concurrent multipath and spoofing, the pooled 3D position root-mean-square error is 3.933 m for MDRNF, compared with 18.684 m for an array-based resilient filter and 21.119 m for a standard deeply integrated extended Kalman filter. These results describe the variation across the 80 independently generated IF signal-level realizations under the specified scenario.