Quantum Magnetometer-Assisted Maximum-Likelihood Fault Detection and Exclusion (QM-MLE-FDE) for Robust GNSS Integrity Monitoring - A Heading-Consistency Framework against GNSS Spoofing

Coherent spoofing attacks against Global Navigation Satellite Systems (GNSS) preserve the geometric self-consistency of pseudorange measurements and therefore defeat conventional Receiver Autonomous Integrity Monitoring (RAIM). Closing this vulnerability requires an integrity channel that is independent of the GNSS signal domain by construction. We propose a Quantum Magnetometer-assisted Maximum-Likelihood Fault Detection and Exclusion (QM-MLE-FDE) framework that uses an optically-pumped or nitrogen-vacancy quantum magnetometer (QMag) as the independent channel. The property we exploit is not the QMag’s picotesla noise floor but its bias-free, calibration-drift-free operation: this is what permits a zero-mean heading residual without a magnetometer bias state, and hence an integrity test that remains statistically independent of the GNSS measurements throughout the deployment lifetime, a guarantee a classical magnetometer cannot sustain. The framework’s central design principle is an initialisation-then-isolation architecture: the QMag is structurally decoupled from the per-epoch measurement loop of a 17-state error-state Kalman filter during steady-state operation (which, as we show at the covariance level, both preserves the heading-consistency χ2 test’s independence and prevents the estimate from tracking the reference and blinding the test), then re-engaged on demand as the primary heading reference once a coherent spoof has been confirmed. A manoeuvre-adaptive heading-accuracy floor and a heading-channel vote into the persistence counter together lower the minimum confirmable spoof bias. Validated on real GNSS pseudoranges and IMU data from the PPC-Dataset (Tokyo; GPS/Galileo/QZSS), the framework achieves sub-10 m RMSE across nominal, coherent-spoofing, and single-satellite-fault scenarios. Under a 500 m coherent spoof with extended GNSS-denied navigation it reduces position RMSE by 55% and time-to-alert by 65% relative to a QMag-disabled ablation, and it lowers the minimum confirmable bias from approximately 500 m to 175 m, a regime the pseudorange-only baseline leaves unmitigated, recovering single-digit RMSE for biases above this threshold. These results establish bias-free quantum magnetometry as a practical foundation for receiver-level PNT integrity. This is a preprint of a manuscript that will be submitted for formal peer review. Copyright may be transferred to a publisher without notice, after which this early version may no longer reflect the final peer-reviewed article.